EP2396679A2

EP2396679A2 - Method for time tracking and positioning of seismic signals of shafts with three components

Info

Publication number: EP2396679A2
Application number: EP10707083A
Authority: EP
Inventors: Charles Naville; Sylvain Serbutoviez; Jean-Claude Lecomte
Original assignee: IFP Energies Nouvelles IFPEN
Current assignee: IFP Energies Nouvelles IFPEN
Priority date: 2009-02-12
Filing date: 2010-02-10
Publication date: 2011-12-21
Also published as: US8898020B2; CA2751108A1; US20120046871A1; WO2010092249A2; WO2010092249A3; FR2942045A1; FR2942045B1

Abstract

The invention relates to a method for tracking the arrival time of seismic waves and to the use thereof for positioning the components of a multi-component sensor. - After acquiring the seismic data using a PSV method, by means of a multi-component sensor, a module signal is built by calculating the square root of the sum of the squares of at least two orthogonal seismic components. Next, the arrival times of a direct seismic wave are tracked on an amplitude extremum of said module signal. Based on said tracking, the seismic components can be positioned in a single frame of reference regardless of the depth of the sensor. For this purpose, a time window is defined on either side of the tracked arrival times, and the azimuth direction is then determined by maximising the energy of the horizontal components within said time window. Finally, the three components are positioned in a single frame of reference, defined relative to said azimuth direction, which is identical for every depth. - The invention can also be used in the exploration or production of an oil reservoir, for example.

Description

POINT-TIME METHOD AND ORIENTATION OF SEISMIC WELL THREE-COMPONENT SIGNALS

The present invention relates to the field of well seismic, and more particularly to the field of seismic data pre-processing techniques acquired by means of multicomponent sensors, obtained during operations of Vertical Seismic Profile (PSV or VSP in English).

The present invention relates in particular to a method for estimating the orientation of a multi-component well seismic sensor. The PSV technique is commonly used for a better knowledge of the structure of the deposit and its neighborhood during the exploration phase, or during the exploitation phase. This technique is also used to define the geological structures in the vicinity of the well to guide the drilling, or to redefine a deviation of the drilling trajectory if the latter has unfortunately not achieved its objective. The most traditional implementation of the vertical seismic profile (PSV) is done by means of a seismic emission made by a surface repetitive source, and a reception made in the well. The latter is carried out thanks to a special well probe, which is parked successively at different depths. The well probe includes a three-component sensor, an anchor system, and a scanning unit for most modern tools. The anchoring system and the mechanical design of the well-receiving probe are such that the three-dimensional recording of the three-dimensional displacement of the formation at the passage of the various seismic waves is faithful to the effective displacement of the formation. This notion of vectorial fidelity characterizes the isotropy of the mechanical anchoring of the sensors of the receiving probe at the wall of the well, and the isotropy sought for the seismic reception in three components. The acquisition can be done with a triaxial sensor placed at a single depth, or by an antenna of triaxial sensors located at adjacent depth receiving levels.

FIG. 1 schematizes generally the acquisition geometry of a PSV 100 conducted in a well 101 drilled substantially vertically, and generally weakly deflected at certain depth intervals. The depth of the well can typically be several thousand meters. The well seismic probe 105 containing the triaxial seismic sensors is lowered by means of a cable 102, which connects it to a surface recording unit 103 from which the field engineer provides all the controls of the well probe. , the remote control of the seismic source 104, as well as the quality control of the measurements. In order to measure, with good vector fidelity, the seismic signal propagating in the rock formation from the surface seismic source 104, this probe is strongly pressed against the wall of the well by means of an anchoring device. symbolized by the arm 106, prior to the recording of each station depth of measurement. In the common configuration of the basic PSV, the so-called "zero-offset" surface seismic source 104 is placed in practice within a radius of 100 m with respect to the wellhead; given the depth of the well, the seismic source 104 is called "surface" even if it is placed a few meters deep. The seismic probe 105 schematized on the figurel may actually represent a receiving probe comprising several levels adjacent seismic reception depth, typically separated by a distance of 15 or 20 meters, each level corresponding to an independent seismic probe, comprising seismic sensors triaxial and an anchoring system.

FIG. 2a shows schematically the propagation geometry 200 of the direct seismic waves emitted by the surface source 206, received by the sensors 201 to 204 located in the well 205: it can be seen, on this diagram, the narrowness of the solid angle 207 (β) propagation in direct arrival. This angle encompasses the entire deep portion of the well in which the PSV measurements are made with a fine regular interval, commonly 15m. In fact, the narrowness of the solid emission angle guarantees, for reception in the well, the hypothesis of waveform constancy of the eigenmodes of substantially homogeneous stratified layer seismic propagation for adjacent depth measurement levels. . Moreover, it can be seen from the diagram of FIG. 2a that the difference in propagation direction is extremely small between the direct spokes, associated with adjacent well measurement levels, such as, for example, the direct spokes a 1 and a 2 associated with the pair. of positions 201-202, or the direct spokes a3 and a4 associated with the couple 203-204. In practice, the immediate proximity of the propagation paths between a common source and substantially adjacent receivers, around ten to a hundred meters per example, is also verified for a down-wave more complex than a direct wave, for example a converted P-type wave S at the right of an interface located at an intermediate depth, for example in the upper half of the total depth of the well . Figures 2b and 2c illustrate two propagation geometries for which the direct arrival of pressure P arrives orthogonally to the direction of the known spatial direction Z component. Consequently, all the P wave energy is recorded by the X and Y sensors of unknown spatial direction, and it is therefore not easy to point a remarkable phase of the direct wave on the raw signals coherently over the Adjacent depth measurement levels.

FIG. 2b schematizes the propagation geometry 210 of the direct seismic waves 212 emitted by the surface source 216, received by the sensor 211 situated in the well 215, and arriving perpendicularly to the well 215. The trajectory of this well is substantially horizontal at the level the position of the sensor 211 whose Z component 213 is axial to the well.

FIG. 2c schematizes the propagation geometry 220 of the direct seismic waves

222 emitted by the surface source 226, received by the sensor 221 located in the well 225, and arriving perpendicular to the substantially vertical component Z 223 of the sensor

221 with three components. The trajectory of this well 225 is substantially vertical at the position of the sensor 221.

The term "multicomponent" qualifies a particular arrangement of a plurality of seismic sensors. For example, a three-component sensor comprises three unidirectional seismic receivers arranged along orthogonal axes, such as geophones or accelerometers. Since seismic waves propagate in three dimensions, a three-component sensor is used to characterize all the seismic waves.

The "component" is the signal from a unit seismic sensor. A three-component sensor generates three electrical signals recorded along three orthogonal axes. In general, the bottom-receiving probe comprises a known axis component arranged either vertically or along the axis of the tool. This axis of the tool substantially coincides with the axis of the well, after anchoring the probe to the well wall, and two orthogonal components, whose exact direction in the orthogonal plane is not known.

It is usual when operating the seismic data, obtained by a method of the PSV type using a three-component sensor, to process only one of the recorded components, in general either the vertical component, the axial component at the well, or the component corresponding to the spatial direction maximizing the energy of the direct arrival pressure wave. Examples of monocomponent treatment of seismic data are described in numerous publications and published works, for example in the following specialized works:

Hardage.B .; "Vertical Seismic Profiling": Principles, Third updated and revised edition; in: Handbook of Geophysical Exploration, Seismic Exploration, Vol.14, 2000, Pergamon, Elsevier Science;

A.H. Balch & Myung W. Lee; Vertical Seismic Profiling: Technical, Applications, and Case Histories, 1984, D. Reidel Publishing Company;

Mari, J. L. et al; "Seismic WeII Surveying", 1991, Technip Publishing, Paris. It is also common to take into account the polarization of the direct pressure mode waves for the orientation and the processing of vertical offset seismic profiles, for which the well is substantially vertical and the position of the source is situated at a greater distance. at 10% of the total depth of the well. We calculate the orthogonal component which maximizes the energy of the direct arrival of pressure wave (P wave) in the orthogonal plane, and in a time window defined by the time point of this direct arrival P. To determine this azimuthal direction in one plane, from two components, one uses a technique of maximization of the energy, for example that described in the following document:

DiSiena, JP, JE Gaiser, and D. Corrigan, 1984, "Horizontal Components and Shear Wave Analysis of Three-Component PSV Data," in MN Toksδz and RR Stewart, eds., Vertical Seismic Profiling, Part B: Advanced Concepts: Geophysical Press, 177-235.

However, the conventional limitation of the treatment to one or two of the components leads to potential indeterminations in the identification of the wave mode of certain arrivals received by the sensor, pressure or shear, on the one hand, and in the positioning geological events on the other hand, which can not be lifted. Time inversion, or the migration techniques of the reflection only seismic events, lead to a plurality of solutions since the azimuthal direction of dip of the reflectors remains unknown. Moreover, when two seismic events arrive at the same time with similar apparent speeds, and appear on some adjacent depth traces (six to twelve for example), it is verified that it is almost impossible to separate them by a conventional monocomponent treatment.

It is therefore essential, to improve the reliability of the interpretation of the seismic data of PSV, to treat the three components. However, the acquisition step does not allow to provide the actual orientation of the cable geophones, while this information is essential to enable data processing. Indeed, the horizontal components from three-component PSV have an unknown and random orientation, because the cables carrying the seismic sensors (geophones) can not control the orientation of these elements. Dealing with all three components may be considered, as the mechanical reception isotropic quality of the PSV probe (summarized as "vector fidelity" of the reception) is good enough. This is the case with most modern PSV tools, in which the ratio between the anchoring force of the part of the probe (or the entire probe) supporting the triaxial sensors, and the weight in the air of the support (or probe) is greater than five. However, although there are accessories for measuring the complete orientation of a tool in a well, such as magnetometers-inclinometers and well gyroscopes, these sophisticated hardware accessories are often not used because they represent a substantial additional cost. In addition, they can deteriorate the mechanical anchoring qualities of the PSV probe to which they are combined. It can be seen that the PSV acquisition step seldom makes it possible to systematically provide the real and complete orientation of the three-component sensors, whereas this information is essential to enable the three components to be processed. In fact, orthogonal components from three-component PSV are mostly unknown and random in orientation. This is particularly the case in the well-inclined well depth intervals, in particular below 10 degrees of vertical inclination, intervals in which the partial orientation devices of the mounting type of the gimballed sensors, or the addition of a pendulum sensitive to gravity measuring the angle of "relative bearing" in the plane orthogonal to the axis of the tool, are rendered inoperative. Thus, in order to completely, efficiently and beneficially treat all the signals from multicomponent sensors constituting a usual set of well seismic data, it is necessary to orient the geophones of the multicomponent sensors.

STATE OF THE ART To determine an azimuthal direction in space from the three components, the covariance matrix analysis technique described in the following documents can be used:

Benhama, A., Cliet, C, and Dubesset, M., 1988, Study and application of spatial directional filtering in three component recordings: geophysical propecting, 36, 591-613, Cliet, C., and Dubesset, M., 1987: The Parameterization of Particle Trajectories, French Petroleum Institute, Report No. 35080.

This technique assumes that the calculated polarization direction belongs to the vertical plane containing the source and receiver positions. This is realistic in a layered sedimentary environment, even with low values dip medium G ^'up to about 20 ° and dipping all, if the source and the sensor well are located in the local structural dip of plane in the vicinity well). This process uses the point of the direct pressure wave performed on the vertical or axial component at the well, whose signal form is consistent in the traces of adjacent depths. This process becomes inapplicable if the direct arrival energy of pressure is too low on the orthogonal components.

This method of estimating the orientation of the three-component sensors by maximizing the energy of the direct P-wave signal in a time window defined by pointing the vertical or axial component to the well, applies well to the geometries of acquisition as offset PSV, which comprises at least one surface seismic source positioned fixed at a certain offset distance from the well (typically equal to 0.2 to 1.5 times the total vertical depth of the well). Likewise, the three-component sensors of the 2D or 3D walkaway configurations, and of the walkaround, which consists of recording on a fixed sensor antenna 3C in the well, the signal transmitted from a plurality of surface source positions is oriented. according to a provision that determines the type of well seismic study. Thus a 2D walkaway corresponds to an online layout of the source points, a 3D walkaway corresponds to a grid of source points, more or less regular, and a walkaround corresponds to a circle of source points shot around the well.

For example, a method of orienting orthogonal sensors using the P-wave forward bias is well illustrated in the configuration of a plurality of source points placed at various azimuths around the well in the document. next :

P. N. Armstrong, "Method of estimating the bearing of a borehole receiver", July 26, 2005, US Pat. 6,922,373 B2

In all these cases, the plurality of surface source positions makes it possible to find several for which the P-wave direct arrival maximization process is well adapted to the precise and redundant estimation of the orientation. . The bottom tool does not need to have a component orientation device.

With redundancy of source points, the adoption of the common propagation hypotheses in the vertical source-receiver plane, and provided that the rectilinearity of the direct P-wave arrival is correct for a sufficient proportion of source points, it 'is not even necessary to know the trajectory of the well, as evidenced by the following document:

Stewart A. Greenlagh and Ian M. Mason, "Orientation of a geophone triaxial downhole", 1995, Geophysics, VOL.60, NO4, p 1234-1237.

Finally, two methods are also known for determining the orientation of the two horizontal components when there is a plurality of positions of sources on the surface, by the document:

X. Zeng, G. A., McMechan, "Two Methods for Determining Geophone Orientation from PSV Data," Geophysics, Vol. 71, No. 4, p. V87-V97, 2006.

A first method, based on the plane of polarization, makes it possible to determine the orientation of the horizontal components of cardan-mounted sensors, from the polarization energy of a time window around the direct arrival P, with indeterminacy of π on the angle found. It is well known in this document that this indeterminacy of π is maintained, whereas it could be easily raised by pointing a remarkable phase of the arrival signal P on the vertical component, in particular an extremum of amplitude, and in imposing a systematically identical polarity on the horizontal component output of the amplitude maximization process, as is practiced industrially. Naturally, this classical polarization method, called PPDI, gives satisfactory results only if the P wave energy is substantial in the horizontal plane, and this method makes use of the massive plurality of surface source points to improve reliability. of orientation and to lift the indeterminacy of π.

In the second method, called RADI, we calculate the relative azimuth between two adjacent depth stations of geophones, by a method of maximizing a correlation between two-component vectors, calculated over a period or a time window around the direct P wave. high energy on the horizontal components. In practice, the method

RADI does not give good results on the real data, that is why it is generally applied after the first method PPDI to lift the indetermination of π mentioned previously.

It is also possible to use the polarization properties of the direct seismic waves P to reorient the components located in the plane perpendicular to the axis of the well, in the case of a particular geometry: acquisition geometry in a deflected well of trajectory known, with a single surface seismic source placed offset of the wellhead, and with registration by triaxial sensors fixed relative to the body of a tool not provided with an accessory for measuring the relative bearing angle . A procedure for maximizing the arrival P on the orthogonal components is then applied. It is also assumed that the direct wave radius P is included in the vertical plane containing the surface source and the bottom sensor. Naturally this reorientation is valid only for the current logging, and it is easy when the direct arrival P is of substantially linear polarization, descending or rising refracted, and not interfered. This technique, well known to those skilled in the art, is described for example in the following document:

M. Becquey and M. Dubesset, "Three component probe orientation in a deviated well" Geophysics, 1990, vol. 55 No. 10, p. 1386-1388.

This orientation method generally provides either two solutions, a double solution, or no solution. In this case, we remove the indetermination of a double solution of P-wave direct arrival radius, keeping only the one that is closest to the line defined by the source and the receiver. If there is no solution, we keep the double solution as an approximation: (Φ = Φ ₀ in equation 6, page 1387 of the previously cited document). This reorientation technique has been used successfully on several real cases, as evidenced, for example, by the illustrations on page 420 of the following document:

C. Cliet, L. Brodov, A. Tikhonov, D. Marin and D. Michon, "Anisotropy survey for reservoir definition", Geophys. J. Internat., 1991, 107, 417-427.

A limitation to all the aforementioned orientation processes taking into account the direct arrival of wave P ₁ occurs when the energy of the projection of the direct arrival on the two non-axial components, or non-vertical, is very low. . This is for example the case if the well is vertical or weakly deflected, in particular with the very current acquisition geometry known as "zero-offset" PSV, for which the surface source located near the wellhead is activated in a unique position, on land or at sea, with a geological structure with local dips of any kind and often unknown. It is known to overcome this limitation by activating an additional seismic source at the zero-offset source, located at a sufficient distance from the well, and preferably in the general azimuthal direction of the geological structure in the vicinity of the well. But this alternative solution is rarely used because it entails additional expenses for the measurement operation as well as an extension of the acquisition period, and therefore of immobilization of the rig at the site. The activation of this additional source is carried out successively or simultaneously with the zero offset source, with the same depth of anchoring and measurement position as the PSV tool, whose three components are to be oriented. In addition, these palliatives with previous methods or geometric configurations of fire, have the disadvantage of not being always applicable, either because of rugged relief for example, or because of material or financial unavailability of additional seismic sources, or to cause of interference from the direct arrival of P-wave by a refracted or secondary diffracted arrival. When it is absolutely necessary to know the orientation of the triaxial sensors, it is advisable to consider the implementation of an additional well offset source, and to compare this solution with the alternative choice of a well tool. combinable with an orientation accessory. But in practice, these two PSV measurement modes are rarely implemented.

Moreover, after orientation of the triaxial components, some multi-component PSV processing programs are limited to the two components contained in the vertical plane containing the sensor and the source, such as the method described in the following document:

C. Esmersoy, "Velocity estimation from offset VSPs using Direct P and converted SV-waves", POS6.4, SEG Abstracts 1987, p538-541.

However, as previously explained, the conventional limitation of the treatment to one or two of the components leads to potential indeterminations in the identification of the wave mode of certain arrivals received by the sensor, pressure or shear, on the one hand, and in the positioning of geological events on the other hand, that it is not possible to lift.

Thus, in order to completely, efficiently and beneficially treat all the signals from multicomponent sensors constituting a usual set of well seismic data, it is necessary to orient the geophones of the multi-component sensors.

The object of the invention is an alternative method of spatial orientation of geophones of a multi-component sensor to overcome the difficulties of orientation of the prior art. The method allows geophones to be orientated in a locally coherent and substantially unique coordinate system for all levels of measurement, unknown to a constant rotation, and then to orient the three components in a geographically-related geographical directions, in order to allow the isotropic treatment of the three components, for the benefit of structural and geological interpretation. The method based on an original point-time technique of the arrival times of a direct wave (P or S) on one of the most energetic phases of a modulated signal.

The method according to the invention Thus, the invention relates to a pretreatment method of seismic data acquired by means of a seismic survey method of the vertical seismic profile type. This seismic survey method of the vertical seismic profile type comprises a seismic wave emission and a reception of these seismic waves by means of at least one multi-component sensor positioned within a well and which is parked at at least two depths. This sensor comprises at least three orthogonal geophones recording as a function of time a first seismic component in a known vector direction, and at least two other seismic components in two directions orthogonal to this known vector direction. The method comprises the following steps: a- a new signal is constructed by calculating the square root of the sum of the squares of at least two orthogonal seismic components, called the modulated signal, and the arrival times of a direct seismic wave on an amplitude extremum of said module signal; b is oriented said seismic components in a single frame regardless of the depth of said sensor, by means of the following steps repeated for each depth:

a time window is defined on each side of said arrival times;

an azimuthal direction is determined by maximizing an energy of said seismic components orthogonal to said known vector direction within said time window;

these orthogonal seismic components are oriented at 360 degrees to said known vector direction in a single reference frame defined with respect to the azimuthal direction which is identical for each depth.

According to one embodiment, the modulated signal can be constructed by calculating the square root of the sum of the squares of the two seismic components orthogonal to the known vector direction. We then point the arrival times of a downward shear wave. According to another embodiment, it is also possible to point the arrival of a pressure wave. Shear wave and pressure velocities can then be calculated from these arrival times. We can deduce reports of speed and / or a Poisson's ratio. According to another embodiment, the modulated signal can be constructed by calculating the square root of the sum of the squares of the three seismic components, and the arrival times of a direct pressure wave are plotted.

Preferably, prior to step a), the signal isotropy is preserved in three components, respecting amplitude ratios and phase differences between the seismic components.

According to the invention, it is possible to improve a signal-to-noise ratio of the three components of the raw signal, prior to the calculation of the module, by means of an isotropic deconvolution of the three components by a single downward pressure wave signal extracted from the component vectorial direction. It is also possible to filter the modulated signal so as to suppress low frequency components, before pointing the arrival times of the direct seismic wave.

According to one embodiment, it is possible to determine the geographical orientation of this unique marker.

The geographical orientation of the single marker can be determined by mounting the multi-component sensor on a double gimbals system, this system making it possible to direct the seismic components by gravity when the inclination of the well reaches a value of at least about 10. degrees.

A well measurement tool including the fixedly mounted multi-component sensor can also be lowered into the well. The geographical orientation of the single marker is then determined by means of a system for measuring a "relative bearing" angle mounted on the measuring tool, which makes it possible to find the orientation of the multi-component sensor when the Shaft tilt reaches a value of at least about 10 degrees.

According to one embodiment, a well measurement tool comprising a plurality of multi-component sensors placed at adjacent measurement depths is lowered into the well, and the geographical orientation of the single coordinate system is determined by coupling at least one multi - component sensors to a geographic orientation measuring tool, such as a magnetometer - inclinometer or a gyroscope.

According to another embodiment, at least a portion of the well is horizontal and the multi-component sensor is fixedly mounted in a well measurement tool. We then determines the geographical orientation of the single coordinate system by assimilating a direction of maximization of a direct pressure wave to a line connecting a position of the sensor to a position of a source emitting the seismic waves.

According to the invention, it is also possible to determine, on different portions of the well, unique markers having a common axis, these portions having overlapping areas making it possible to determine an angle of rotation to be applied to these single markers, so as to obtain a single benchmark for the entire well.

Finally, the orientation can be used in a single reference frame in an automated way to obtain a quality control of the seismic components, immediately after the acquisition of measurements in the field.

Other characteristics and advantages of the method according to the invention will appear on reading the following description of nonlimiting examples of embodiments, with reference to the appended figures and described below.

Brief presentation of the figures

FIG. 1 schematizes generally the acquisition geometry of a PSV with surface source and probe comprising a single level of seismic reception, lowered into a substantially vertical well by means of a cable. FIG. 2a illustrates the narrowness of the direct arrival propagation solid angle, which guarantees the hypothesis of waveform constancy of the homogeneous seismic propagation modes in stratified medium, substantially homogeneous for adjacent depth measurement levels. .

FIGS. 2b and 2c illustrate two propagation geometries for which the direct arrival of pressure P arrives orthogonally to the direction of the known spatial direction Z component; this situation can occur in a horizontal well (Figure 2b), in a deviated well, or in a vertical well (Figure 2c), when the source is well away from the well.

Figures 3a and 3b show the arrangement of the three components of a triaxial sensor mounted on double gimbals of the "turret" type, and placed in a deflected well: vertical plane tangent to the well (Figure 3a); horizontal plane (fig.3b).

Figures 4a and 4b show the arrangement of the three components of a triaxial sensor fixedly mounted in the tool, in the plane orthogonal to the axis of the well and the probe (fig.4a), which allows to illustrate the angle of "relative bearing", and in the vertical plane tangent to the well (fig 4b).

FIG. 5 illustrates the spatial attitude of the trihedrons of the seismic sensors in the vertical and deviated parts of a well before orientation (a): fixed mounting, c): mounting on double gimbals), and the configuration of the trihedrons after orientation (b). ).

FIGS. 6a, 6b, 6c and 6d illustrate the principle of propagation of the seismic volume shear waves (S waves) in a substantially homogeneous medium, and the orientation method of the two non-oriented orthogonal components in an intermediate reference frame. unique according to the invention. - Figure 7 is a flowchart of the entire orientation process according to the invention.

FIG. 8 illustrates the particle motion of the down-wave S in the orientation calculation time window, in the reference of the tool; the angle AG calculated and corresponding to the arrow superimposed on the hodogram indicates the maximization azimuth. FIGS. 9a, 9b and 9c represent the isotropic gross rejects of the respective components Z ₁ X, Y measured by the well tool, versus the increasing depth from left to right.

Figures 10a, 10b, and 10c show the isotropic raw reps of the respective vertical, H-North (HN), and H-East (HE) oriented geographic components of the tool, versus the increasing depth from left to right, and the coherence obtained on the arrival S in the rectangular window, illustrates the efficiency of the orientation method of the two orthogonal components according to the invention.

DETAILED DESCRIPTION OF THE METHOD The invention relates to a method of specific pretreatment of seismic data acquired during the implementation of a seismic survey operation of vertical seismic profile type. According to one embodiment this operation comprises a single surface position for the emission of seismic waves, located near the drilling apparatus, and a reception of the seismic waves by means of a multi-component sensor positioned within from a vertical well to weakly deflected. The sensor (mounted in the seismic probe) has three orthogonal geophones. And this sensor is parked at least two depths close to a few meters. The signal recorded in the axial direction of each unit sensor is called the component of the signal. The receiving probe records as a function of time at least one component in a known vector direction, vertical or axial to the measurement well, and two so-called "orthogonal" seismic components, that is to say orthogonal to the known vector direction component.

Figures 6a and 6b illustrate the basic physical principle of the propagation of a shear seismic wave, called S (shear). There are two types of volume waves: pressure waves (P waves), whose polarization, or direction of movement of the formation at the passage of the wave, is parallel to the direction of propagation, not shown in the figure 6a, and transverse waves (S waves) polarized substantially perpendicular to the direction of propagation, illustrated in Figure 6a.

FIG. 6a represents a diagram of a vertical plane of propagation 600: the surface 601 symbolizes not only the surface of the ground but also the first heterogeneous ground layers up to about 200 meters thick, or even the rough surface of the bottom of the sea, which are the seat of P-mode mode conversions in S mode. All wave modes propagating in depth along the vertical axis 602. Thus, from a shallow depth of ground, the S wave propagates vertically in a substantially homogeneous medium at the scale of the seismic wavelengths used. The medium is isotropic or with weak anisotropy for the direction of vertical propagation (resulting in an ordinary birefringence), and it includes a possible stratification of layers of variable characteristics, with a possible dip for all layers or not. The medium is thus representative of usual sedimentary formations or basement. The diagram of FIG. 6a illustrates a birefringent medium in which a complex shear wave train propagates vertically in depth along the axis 602 and comprises:

a fast shear wave S-f having any vibration shape 603, linearly polarized in the constant direction 611 orthogonal to the direction of propagation 602.;

a slow shear wave S-s having a vibration form 604 generally different from 603 and linearly polarized in the constant direction 612 orthogonal to both the propagation axes 602 and the fast S-wave S-wave polarization axis 611.

The polarization vectors 611 and 612 symbolize the vibration directions of each of the wave modes S propagated according to the vertical, but also the amplitude of the signal. The amplitude of the vibration signal indiscriminately characterizes the displacement, the speed or the acceleration of the seismic movement of each of the wave modes Sf and Ss, whose direction and waveform remain constant throughout the propagation. according to the physics of the propagation phenomenon. FIG. 6a shows, for successive instants of propagation ti 610 (i = 0.1, 2, i) measured from the moment of activation of the surface seismic source, the depth z _f i 613 reached by the fast wave Sf, and the depth z _s i 614 reached by the slow wave Ss, the foot of the polarization vectors respectively 611 and 612 indicating the exact depth reached by each of the two waves S at time ti 610. In fact, the depth z _f i 613 of the fast wave Sf is progressively and slightly higher by a few percent at the depth z _s i 614 during the spread. FIG. 6b represents, in the horizontal plane 650, orthogonal to the direction of vertical propagation, projection diagrams of the theoretical vibration motion forms of the wave train S. They each represent a continuous succession of the points [X (t), Y (t)] seismic signals, in a time window centered, in this case, around the time ti of the wave arrival S to illustrate: these diagrams are also known under the names of "particle movement" if the signals X and Y represent a displacement at the passage of the seismic wave, or even "hodograms" if the X and Y signals represent a movement speed, measured for example by a geophone-type sensor. They are also called "polarization diagram" or simply "polarization" if the X and Y signals indifferently represent a displacement, a speed or an acceleration.

Figure 6b shows on the right column 653 below the indication "IN", and for successive fixed vertical depths zi (i = 0.1, 2, i) 660, reached by the wave train S in a time window around the propagation instants ti 610 (i = 0,1, 2, i) defined in FIG. 6a, the theoretical polarization diagrams 651 of totally random orientation in the horizontal plane for each station depth zi, that the a well seismic probe used for the recording of PSV is recorded in the reference frame and anchored to the well wall according to a random azimuth. The randomness of rotation of the logging cable at the end of which the well probe is fixed is well known to those skilled in the art, this hazard being transmitted to the fixed-mount seismic sensors in a probe. Moreover, if said receiving probe comprises turret-type gimbals on which the seismic sensors are mounted, the rotation of the gimbals around the axis of the probe is free when the probe is in a vertical well, which also causes an orientation. random horizontal sensors, even if the azimuthal orientation of the probe was known.

FIG. 6b symbolically indicates the process P1 (666) according to the invention, via an arrow between each polarization diagram 651 of the data entered in the "IN" 653 right column, and the corresponding oriented polarization diagram in the left column.

"OUT"663; this process consists in determining the maximum amplitude direction 652 on the chart 651, independently for each well seismic depth of measurement, by a known technique of linear regression of the group of points X (t), Y (t) in an appropriate time-limited window, if possible shorter than the window-time corresponding to diagrams 651 or 661, then to apply a vertical axis rotation to horizontal seismic signals so as to make the maximum amplitude direction 651 coincide with an arbitrary constant azimuthal direction 662 in the diagram 661, at first. This process is well suited to the case of PSVs called zero offset in vertical wells, where the P wave source usually used generates unintentionally and very often in practice energetic shear waves at the passage of the altered and heterogeneous surface area or the rough surface of the bottom of the sea.

FIG. 6b shows on the left column 663 below the indication "OUT", and for successive fixed vertical depths zi (i = 0.1, 2, i) 660, reached by the wave train S in FIG. a time window around the instants of propagation ti 610 (i = 0,1, 2, i) defined in FIG. 6a, the theoretical polarization diagrams 661 that are expected to be observed in a fixed reference of the horizontal plane, for example a geographical landmark (North, East). For example, we take zi = (z _f i + z _s i) / 2, with reference to FIG. 6a. Note that in case of isotropy of the propagation medium for the vertical direction, zi = z _f i = z _s i represents the depth reached by two waves S of any orthogonal polarization. The polarization diagrams 661 of FIG. 6b represent the particle motions of the wave train S illustrated in FIG. 6a, for the same succession of propagation instants ti: it may be noted that the successive diagrams on the left column 663 are of very similar shape, with a maximum direction of amplitude 662 identical. The differences in rotundity observed typically reflect the azimuth anisotropy of birefringence, resulting from the small difference in speed between the fast S-wave Sf and the slow S-wave S. Given the time length of the seismic wavelets 603 and 604 (FIG. Sf and Ss eigenmodes emitted almost simultaneously, the shape of the diagrams in Figure 6b is elliptical and reflects the interference of the two S wave modes that can unfortunately not be distinguished visually in practice in the general case on this genre of diagram. It has been found by experiment that an individual energy seismic arrival in P wave of linear polarization or in S wave of any elliptical polarization, P or S waves whose polarization form remains substantially constant during propagation, shows a coherence of time amplitude extremes from one depth to the next, and in particular, that the shape of the modulated signal remains substantially constant for a high energy down-wave. Publications showing hodograms of direct S-wave trains emitted by a source S placed at a short distance from the well, and reoriented in a fixed geographical reference point by means of offset and activated P-wave source can easily be found. in the same measurement run, and indicating that the hodogram shape is substantially stable throughout the propagation along the vertical well and in a horizontal stratigraphic medium having significant S-wave azimuthal anisotropy, for example in the documents following: Charles Naville, "Detection of Anisotropy Using Shear-Wave Spitting in VSP Surveys; Requirements and Alpplications", SEG Expanded Abstracts, 56 ^th int. SEG meeting, 1986, Houston, S5.2, pp. 391-394. lan Bush and Stuart Crampin, "Paris Basin VSPs: case history of anisotropy anisotropy and crack anisotropy case of thin layer (or lithology). Journal Int., 1991, No. 107, pp. 433-437.

Nicoletis, L., Cliet, C. & Lefeuvre, F., "Shear-wave Splitting measurements from multishot VSP data, Expanded Abstracts, 58 ^th int., SEG meeting, 1988, Anaheim, POS 6.1, pp.527-530; the three aforementioned documents, the hypothesis of the polarization fixity of the eigenmodes (FIG. 6a) for a given direction of propagation is admitted explicitly or implicitly, and it can be visually verified that for a vertical propagation with a small deviation of the vertical, the shape of the polarization pattern remains similar with the propagation, as well as the azimuthal direction of maximum amplitude Thus, even in the presence of birefringence anisotropy without differential attenuation of drastic amplitude between the two wavenumbers S, it is easy to verify mathematically that the maximum amplitude direction of a two-component signal of a non-interfered direct waveform S remains substantially fixed along a direction of propagation. This is shown schematically in the left column of Figure 6b. Thus, by determining the azimuthal direction for each depth at which a record of the horizontal components is available, a single coordinate system is defined. This method is all the more precise in that (1) the mechanical coupling of the single-level receiving probe or of each of the receiving probes of a multi-level well tool ensures a good vector fidelity of the measured signal, and that (2) the surface source remains in fixed position and always emits the same signal form during the entire PSV operation.

The method includes an orientation of the seismic components in a locally coherent coordinate system regardless of the depth of the probe. Fig. 7 is a flowchart 700 of an embodiment of the orientation method according to the invention. It is a question of orienting the two orthogonal components to a vertical component, or to an axial component to the well. The method relies on the analysis of downward S waves to define a time window necessary for the determination of a rotation angle. Indeed, the recorded P waves usually having too little energy on the horizontal components of PSV, they do not allow to reorient these components. This is particularly the case for PSVs without offset, carried out with wave sources of pressure in wells whose trajectory is close to the vertical. On the other hand, the energy of the direct waves S or waves converted from P to S during the downward propagation is very often sufficient to apply the method.

Thus, according to one embodiment, the method comprises the following steps for each measuring depth station: in block 701, the two orthogonal components to be orientated, that is to say that one applies, are isotropically deconvolved the same operator, at the same time on both components. This operation is only considered if the result of operations 702 to 704 is not satisfactory. in block 702, the module, denoted M (t), of the two raw components to be oriented, or an exponential power (n) of the latter, is calculated; in block 703, the module M (t) is filtered so as to eliminate its low frequency components and possibly high frequencies containing only noise, so as to facilitate the following operation. in block 704, the time Tp of a remarkable phase of the signal of the filtered module obtained above is pointed out. For example, a peak or a dip of amplitude is pointed out, possibly refined by application of an industrial algorithm of correlated or semblance pointing.

Operations 702 to 704 may be automatic, cascadable, and in one pass, but may result in unreliable or inaccurate scoring. In such a case, it is possible to modify the filter of the block 703, or to envisage previously applying operations 702 to 704 the operations described in block 701 intended to increase the signal-to-noise ratio of the wave S that the it is desired to point, all arrivals interfering with the desired arrival S being considered as noise. in block 705, a time window is then defined around the arrival arrival time S, of a constant length for all the measurement depth levels and at least equal to half a period of the filtered module signal, or even a greater length in the case of an interference wave S or a low signal-to-noise ratio in the block 705, the azimuthal direction is determined, by maximizing the energy of the input components X (t), Y (t), raw or filtered to retain only the highest signal to noise ratio frequencies, according to a known industrial process. The raw input components X (t), Y (t) are then rotated in the plane orthogonal to the gross component Z, in a single intermediate coordinate system defined by the azimuthal direction calculated previously. We make sure that the amplitude of the signal output in the maximization direction remains of identical sign, for example positive, at the pointed time Tp resulting from the operation 704, for all the measurement levels of the PSV. in block 707, the components are calibrated in a single intermediate reference, defined with respect to a geographical reference, or a reference linked to the known trajectory of the well if additional information is available for this purpose. in block 708, the necessary rotations are applied to the single intermediate marker to restore the three components of PSV in a geographical reference linked to the terrestrial globe, according to a known procedure.

Each of the preceding steps is described below: A. Pointed

Block 701: possible preliminary deconvolution: in some cases, the signal of the S-wave forward train is apparent over a large recording time interval, it is potentially interfered with by other relative energy waves more low, but not negligible, which has the effect of making the point of the filtered module more imprecise. It may be useful in such cases to apply a multichannel isotropic deconvolution operation, identical for the two "horizontal" components (components orthogonal to the substantially vertical direction) of each of the depth of measurement levels, and identical in several depth levels. This makes it possible to reduce the length of the wave train S on which the arrival time is to be pointed: the deconvolution can be carried out by extracting the wave signal P on the vertical component, in order to deconvolute a converted wave PS on the two horizontal components, for example according to the method described in the patents of Nigel Anstey GB 1, 569 581 of 27-09-1977 or CA 1, 106,957 of 9-12-1977 titled "Seismic delineation of oil and gas reservoirs using borehole geophones ". Deconvolution can also be carried out simply with the existing isotropic and multichannel industrial algorithms for surface seismic or for well seismic, Wiener-type or even frequency-spectrum balancing, both based on the amplitude spectrum of the sum of the autocorrelation signals of each of the orthogonal horizontal components to be deconvolved, taking into account the invariance of this sum of autocorrelation signals with respect to the orientation of the two horizontal input components. • Block 702: Calculation of the module signal M (O of a two-component signal, and invariance.

After making sure that the basic preprocessing operations of the PSV ₁ unit records such as editing, vertical summation, potential source amplitude normalization prior to summation ..., have been performed isotropically the modulated signal, denoted M (t), which represents one of the polar coordinates deduced from the two raw signals in Cartesian coordinates X (t) and Y (t, d) is calculated as follows:

M ² (ή = X ² (ή + Y ² (ή, for all times t

If the PSV tool, whose wall coupling is mechanically isotropic, is anchored at a given depth with an azimuthal direction different from an unknown az angle around the vertical shaft axis, the tool records the horizontal components X1 (t) and Y1 (t) which are expressed as follows as a function of X (t, d) and Y (t, d):

X1 (t) = X (t) .cos (az) + Y (t) .sin (az) Y1 (t) = - X (t) .sin (az) + Y (t) .cos (az) II is easy to see that for any value of the angle az:

The module remains identical regardless of the orientation of the sensors associated with the components X (t) and Y (t), it is invariant with respect to the rotation, always positive value. The module of a two-component signal is also called "M2" in the remainder of the description. The module of a three-component signal, called "M3", defined by:

M ² (t) = X ² (t) + Y ² (t) + Z ² (t), is also an invariant with respect to any spatial rotation.

It is interesting to calculate the M3 module when one wants to point a direct pressure wave (P) whose direction of arrival is orthogonal to the well in certain configurations of the acquisition geometry. For example, on some intervals of the measurement well close to the horizontal, when the source is located close to the plumb of the seismic sensor 3C, as illustrated in FIG. 2b, or else in the Offset-type acquisition configurations. VSP and walkaway, at some well near vertical intervals, when the surface seismic source is sufficiently far from the well, so that the seismic ray arrives at the well with a horizontal incidence, as shown in Figure 2c. As this arrival P pointing procedure is valid whatever the incidence of the radius, it automates the pointing of the direct P wave by pointing the module M3, in particular for walkaway 2D and walkaway 3D studies, this latter configuration also known as 3D-VSP.

One can also choose to work on an exponential power of the signal M (t), in order to amplify the variations of amplitude of this signal. The advantage of the property of invariance of the module is to allow precise pointing operations of the time of a remarkable phase, related to a particular moment of the wave arrival train S _1, for example a well-individualized local extremum, without knowing the prior orientation of the constituent signals of said module.

Block 703: Filtering of the module Mit) According to a preferred embodiment, the method comprises a filtering of the module

M (t), so as to remove its low frequency component and make this signal more readable. For example, one can typically use a 5-60Hz bandpass filter for PSVs.

Block 704: Time stamp of a particular extremum of the signal of the filtered module.

The point of time Tp of an amplitude extremum of the module, preferably filtered, is described in relation to FIGS. 6c and 6d.

FIG. 6c shows a horizontal projection 670 of a polarization diagram 671 of a direct feed train S chosen in a wide appropriate time window.

The modulating signal M (t) 672 is represented graphically as one of the polar coordinates derived from the two raw orthogonal recorded signals X (t) and Y (t) in Cartesian coordinates and of arbitrary orientation.

By definition, whatever the time t: M ² (t) = X ² (t) + Y ² (t).

The signal M (t) 672 has the mathematical characteristic of being invariant with respect to the Cartesian reference of the raw signals X (t) and Y (t) measured, and in which the module is calculated. Likewise, the shape of the polarization diagram 671 in the time window considered is independent of the marker, with one rotation. The raw signal vectors X (t) and Y (t), as well as the module vector M (t), have the same origin 673 (zero of the amplitudes), the amplitude of the vector M (t) always being positive or zero.

Given that the polarization pattern of the direct wave S considered is also almost constant during the seismic propagation depth, the method according to the invention allows to accurately determine the time point of a remarkable phase of the signal of S-wave undirected, independently of the Cartesian frame of the raw signals X (t) and Y (t), for example the time of one of the local maxima 674 of the module signal M (t) 672. FIG. 6d shows a schematic example, as a function of time t, of the modulating signal of a wave train S whose energy is greater than that of all the other waves received at the same time by the seismic sensor. It is found, by experiment, that this modulated signal remains substantially identical as a function of the recording depth, with a time offset corresponding to the propagation of the S waves. In practice, in order to amplify the recognition of the local extrema of the modulated signal, the low-frequency components are eliminated by a low-cut filter, the result of which is the filtered signal Mf 693, which also represents the difference between the raw signal 691 and the associated smoothed signal 692. The amplitudes of the filtered signal Mf can also be raised to an exponential power, to further facilitate the pointing by visual method, or by calculation of similarity or correlation between PSV depth measurement stations. The time Tp of amplitude peak 695 is in practice easier to point, unambiguously, on the filtered signal Mf 693 than the peak 694 on the raw signal M 691, both for the eye and for most industrial computer algorithms. of time pointing. The method according to the invention, making it possible to obtain a precise pointing of a remarkable phase of a shear wave signal with two orthogonal components in the plane of polarization without prior orientation, leads to immediate applications: indeed, this kind of pointing makes it possible to access the knowledge of a wave arrival time S at a constant nearly identical for all the stations depth of measurement of the PSV, and consequently to the knowledge of the interval speeds in mode S. By combining the time S with the measurement of the time in wave P generally carried out on the vertical or axial component to the well, one accesses for example the ratio of the interval velocities Vs / Vp and the Poisson's ratio. Young's modulus is also available if the formation density is known elsewhere. It is also possible to point an S-wave at data from dipole or quadrupole wave-S ultrasonic wave log measurement well tools having flexural wave sources and receivers, without the need to know the waveform. orientation of the tool. This can lead to design simplifications and reduced operational cost since the hardware elements for orientation measurement are no longer needed.

B. Orientation in a single, coherent landmark

Block 705: Determination of the azimuthal direction of maximization of the energy of the two raw components to orient and rotation of the input signals in a coherent intermediate reference frame. The definition of a calculation time window is defined by a constant time difference of the order of 10ms to 20ms on either side of the pointed time Tp on the filtered module, so that the time window includes at least one half period of the dominant period the arrival pointed. The invariance of the modulating signal with regard to the orientation of the sensors causes the coherence of the pointed time of the arrival S, and consequently guarantees its validity for subsequent uses or to know the wave speeds S as a function of the depth. .

The raw signals X (t) and Y (t) are taken for each of the measurement depths of the PSV, which is optionally filtered by cutting off the high noisy frequencies. Then, we search for the azimuthal direction that maximizes the seismic energy in the plane of the two input components and in the previously defined time window, using a common technique of maximizing energy, such as that described in the aforementioned documents:

Benhama, A., Cliet, C, and Dubesset, M., 1988, Study and Application of Spatial Geophysical Prospecting, 36, 591-613,

The component corresponding to this direction of maximization is denoted Hmax (t), the calculated angle between Hmax (t) and the first component X (t) is denoted amax180 and is known only to 180 °. This indetermination is raised by choosing, for example, that the amplitude of the output component Hmax is imperatively rendered positive sign at the instant of the pointed time Tp for any depth of measurement, by proceeding as follows: if Hmax (Tp) > 0, then we define an angle amax360 = amax180, expressed in degree if Hmax (Tp) <0, then we define an angle amax360 = 180 + amax180

FIG. 8 shows hodogram type 800 polarization diagrams drawn in the reference 801 of the gross horizontal X and Y components of a real PSV, recorded in a vertical well with a source with a very small offset, and a well tool comprising three fixed sensors without orientation accessories and whose mechanical coupling ensures good vector fidelity. To the left of each of the hodograms is a legend indicating the depth 802 of measurement PSV, the times in millisecond of the start 803 and the end 804 of the maximization calculation window 55 ms which follows the dotted time done beforehand on the filtered module, the maximum value 805 of the amplitude of the signal vector 807 calculated in the angular direction of maximization AG 806 expressed in grade (GR) from the reference component X 801 of the well tool and in the opposite direction of the Clockwise. The angle 806 noted AG in FIG. 8 corresponds to the angle α max 360 defined above modulo 360 degrees or 400 degrees.

The subsequent rotation of the raw components X (t) and Y (t) of angle amax360 which is applied over the entire length of the recorded signal, makes it possible to obtain output signals in a single reference frame which makes each of them coherent as a function of the depth. The angle amax 360 can be added with a possible constant.

Figures 9 and 10 show the three component PSV signals before and after orientation. FIGS. 9a, 9b and 9c represent the isotropic raw reefs 900 of the respective vertical Z and horizontal X ₁ Y components of the well tool, as a function of the increasing measurement depth ("Measured Depth") MD 901 from the left to right. The replay is called "standardized isotropic 3C", indicating that a constant gain has been applied identically to the amplitudes of the three components, but varies according to the depth, so that the amplitude of direct wave arrival P on the vertical component Z is identical to any depth. The direct arrival of P wave 902 is almost invisible on the horizontal components X, Y on which the wave time P has been represented by a line 903. A direct arrival of wave S 904 is clearly identified on the horizontal components X, Y by its slope greater than that of the P wave 902-903, and one observes coherence defects of the waveform S in the rectangle 905. These defects are associated with the random orientation of the horizontal sensors and the direct arrival maximization direction S illustrated in Figure 8.

Figures 10a, 10b and 10c show the normalized isotropic repeats 1000 3C of the respective vertical Z oriented components, HN and horizontal HE oriented in the respective geographical directions North and East, as a function of the increasing MD 1001 depth from left to right, with the same scales of time and depth as those of Figures 9a, 9b and 9c. The four levels of shallow measurements are missing. The time of the direct arrival of wave P 1002 was represented by the line 1003 on the horizontal components HN and HE. The coherence of the direct wave arrival S 1004 on the horizontal components HN and HE in the rectangle 1005 is much better than in the corresponding rectangle 905 of FIGS. 9a, 9b and 9C, which confirms the good orientation obtained.

At this point, the three components are oriented 360 degrees in a substantially unique frame. This landmark is consistent for each depth. This pretreatment makes it possible to carry out a treatment of the three isotropic components, even if this marker is of unknown azimuthal direction. C. Orientation in a geographical landmark

Block 706: Calibration of the intermediate coherent reference with respect to the geographical reference:

It is desirable, when possible, to additionally orient the horizontal components in a single coordinate system of known geographical orientation. To do this, it is necessary to determine the geographical orientation of the single intermediate marker obtained at the output of the operations of block 705 of FIG. 7. This azimuthal calibration operation of the single marker enables the geological interpretation of the treatment results of the PSV at three subsequent components, such as that shown in US Patent 6,076,045, focused on the dip and azimuth determination of seismic reflectors.

Several methods for calibrating the unique reference system may be used: a) For example, the residual energy of the P-wave input, which is sometimes greater on the horizontal components of the shallower measurement levels of the PSV, may be used. , making the classical assumption that the polarization of the direct P wave is in the azimuthal direction of the segment which connects the position of the source and that of the sensor. This has been done to obtain Figure 10, by performing an additional rotation of constant angle with respect to the azimuthal direction of maximization of the direct arrival S, so as to orient the horizontal components in the geographical coordinate of FIG. 10 shows. Indeed, the direct wave arrival P 1003 has a vertical incidence of the order of 10 degrees on the shallower levels situated between 1000 and 1100 m for the PSV data shown in FIGS. 9c (Metered Depth 901) and Figures 10a-10c (Signals on the right side of the figures) b) Alternately, directional measurements of the three components can be used by various instruments or instrumentation accessories. complete or partial orientation, if they have been descended in a coupled way to the single-level PSV tool.

The orientation is called complete, when all the parameters allowing the orientation (angles of "Relative Bearing", of Vertical deviation of the well and Azimuth of the deflected well) are measured on all the levels depth of measurement of the PSV. This is possible with a gyroscope type tool coupled to the mono-level PSV tool for example.

The orientation is said to be partial, if the orientation measuring tool is coupled to at least one of the measurement satellites of the PSV seismic tool, if the latter comprises several depth levels measured simultaneously. The orientation is also called partial, if the orientation measurement is limited to a given depth interval (for example the limitation to the open-hole well, not jacketed with metal tubes for a tool detecting the direction of the magnetic north), or at a range of deflection angle of the well (such as inclinometer-type devices, measuring pendulum Relative Bearing and universal joints, sensitive to gravity, and rendered inoperative for small vertical deviations from the well).

• block 707: Rotations of the seismic signals between the intermediate coherent reference and the geographical reference, when the orientation of the tool is partial.

In order to facilitate the understanding of the invention and its object, a brief overview of the known material means of orientation of seismic and non-seismic well tools is set out below:

B Comprehensive and Accurate Guidance for a High Operating Cost Well Probe: The logging industry has magnetic tracking capabilities that identify the direction of the Earth's magnetic field, if operating in an open hole, often combined with precise inclinometers rendered insensitive to vibrations and able to perform continuous measurements during the ascent of logging tools also in continuous operation. Precise inclinometers make it possible to know the relative bearing of a tool in a tube well from a few vertical inclination degrees of the well, the trajectory and the angles of inclination and azimuth of the well being known elsewhere. Well gyroscopes are also commonly used to accurately measure the trajectory of the well; their use in combination with other logging tools is sporadic, but not uncommon. Using the means described above, the orientation of the components is then perfectly measured in open hole or tube. u Means of partial orientation and imprecise, but inexpensive, a well probe: a) for deviated wells of known trajectory, it is customary to mount the triaxial sensors on double gimbals with architecture called "turret", comprising an axis of rotation parallel to the well axis, and a horizontal axis perpendicular to the vertical plane tangent locally to the well. Figures 3a, b show the arrangement of a triaxial seismic sensor 311 mounted on such gimbals, and placed in a deviated well

310: FIG. 3a shows a projection 300 in the vertical plane tangent to the well 310, which comprises the downwardly directed vertical component Z-down 301, and the horizontal component XH 302 oriented in the azimuth of the increasing measured depths of the well. ; the other horizontal component YH 303 is orthogonal to the tangent vertical plane shown. The vertical inclination angle of the well 304, or deviation, is commonly referred by DEV in the industry. FIG. 3b shows a projection 350 in the horizontal plane in plan view: the trajectory of the deflected well 310 appears as any line, in the geographical reference 320, the horizontal component XH 302 is tangent to the well at the position of the sensor 311, the horizontal component YH 303 is disposed at + 90 ° with respect to

XH302, in top view. The seismic components HE and HN oriented in a geographical reference frame 320 are recalculated from the components XH 302 and YH 303 by rotation of angle HAZI 305 around the vertical, HAZI 305 corresponding to the azimuth of the well locally to the position of the sensor. 311. The angles DEV 304 and HAZI 305 are generally known and measured independently of the operation of

PSV by very accurate measurements of the trajectory of wells made using gyro or magnetometer - inclinometer type means mentioned above. The mounting of triaxial sensors on double gimbals turret type can guide the seismic sensors three components by gravity wells sufficiently inclined relative to the vertical, typically from a threshold value of the order of

10 degrees of the vertical inclination of the well, this threshold may vary from one tool mark to another; in practice, given the friction forces inherent in this kind of mechanical device, the orientation becomes more precise when the inclination of the deflected well increases. For the low well deviation values below the threshold value of about 10 degrees, the orientation of the orthogonal components is not known. While there is similar uncertainty as to the true orientation of the Z-down component relative to the actual vertical direction, this does not significantly alter the processing results or subsequent interpretative conclusions. Alternatively and currently, it is known to mount fixed three-component seismic sensors in a PSV tool, further comprising a device for measuring the Relative Bearing angle in the plane orthogonal to the axis of the beam. PSV tool: Naturally this kind of device commonly called "Relative Bearing sensor" is inoperative in strictly vertical wells and restores a measure of Relative Bearing which is significant only beyond a low value of the vertical inclination of well, of the order of 10 degrees; the measurement of the Relative Bearing becomes more and more accurate as the inclination of the deviated well increases. FIGS. 4a, b show the arrangement of a triaxial seismic sensor mounted in a fixed manner in the tool: FIG. 4a illustrates the definition of the Relative Bearing angle by the angle between the high generator of the cylindrical well and a direction of the PSV tool in the plane orthogonal to the axis of the tool, with a positive sign convention in the clockwise when looking at the orthogonal plane in the direction of increasing curvilinear depths of the well. FIG. 4a shows a projection 400 in the orthogonal plane to the Z axis of the well 410, at the level of the sensor 411, in a view from above, the arrow 412 indicating the direction of the increasing measured depths of the well; the relative bearing angle RB 430 is defined by the angle between the direction XV 422 orthogonal to the axis of the well 410, contained in the vertical plane tangent to the well and pointing upwards, with the reference direction X 419 of the probe containing the sensor 411, corresponding to the measured X 419 orthogonal seismic sensor; the angle RB 430 is positively measured 431 clockwise when looking in the direction of the arrow 412.

FIG. 4b shows a projection 450 in the vertical plane tangent to the well 410 locally at the position of the sensor 411, which comprises the component Z 421 measured by the tool, axial to the well and pointing downwards, and the component XV422 previously calculated in FIG. the direction of the axis origin of the relative bearing angle (RB = O); the horizontal component YH 403 is orthogonal to the tangent vertical plane shown. The vertical inclination angle of the DEV 404 well is indicated between the axis Z 421 axial to the well direction and Z-down 401, vertical seismic component oriented downward; the horizontal component XH 402 oriented in the azimuth of the increasing measured depths of the well and the Z-down seismic component 401 are obtained from the components XV422 and Z 421 by DEV corner rotation

404 around the YH 403 axis.

Three configurations of partial measurement of orientation are considered below:

C1: the mono-level well tool contains triaxial sensors mounted on double gimbals with a so-called "turret" architecture, as illustrated in FIGS. 3a and 3b, and in a restricted depth interval containing at least one PSV measuring station, the Deflection of the well is sufficiently large (at least about 10 degrees) to allow rotation of the gimbals under the action of gravity: the component 301 Z-down is then naturally oriented in the vertical (fig.3a). The horizontal geographical components 320 HN, HE (FIG 3b) are obtained by rotation of the components XH 302 and YH 303 measured, around the vertical, of the known angle HAZI 305 at 360 degrees, corresponding to the azimuth of the vertical plane. tangent to the well at the sensor position.

[HE ₁ HN] = Rot (HAZI). [XH ₁ YH]

C2: the mono-level well tool contains triaxial sensors fixedly mounted in the well tool, as illustrated in FIGS. 4a and 4b, and in a depth interval With at least one PSV measuring station, the deflection of the well is large enough to allow a precise measurement to a few degrees of the relative bearing angle RB 430 illustrated in FIG. 4a: three successive rotations are then applied in this case. order: [XV ₁ YH] = Rot (RB). [X ₁ Y] ₁ rotation in the plane orthogonal to the axis of the well, then [XH, ZV-down] = Rot (DEV). [XV, Z], rotation in the vertical plane tangent to the well at the position of the well tool, as shown in fig. 4b, then [HE ₁ HN] = Rot (HAZI). [XH ₁ YH], rotation in the horizontal plane, as shown in fig. 3b. C3: the well tool has a plurality of receiving probes placed at adjacent measurement depths each containing triaxial sensors fixedly mounted in the well tool; in addition, one of the probes is combined with a complete tool for measuring orientation. In this configuration, after rotation of one of the orthogonal components to the axis of the well in a single reference system, the difference between the depth of measurement and the orientation measurement tool is calculated for all depth stations measured with the probe. previous angle of rotation and the relative bearing angle measured, then the value of this difference is interpolated for the adjacent depth levels which do not benefit from orientation measurement; the interpolated difference angle obtained is the "relative bearing" angle RBi to be used for the rotation of the orthogonal components of the intermediate marker. The three rotations described for the C2 configuration above are then applied, taking the relative bearing angle RBi for the first of the three rotations.

FIGS. 5a, 5b and 5c illustrate, in the vertical projection plane 500, the spatial attitude of the trihedrons of the seismic sensors 511 to 513 and 521 to 523 in a well 510 comprising a vertical part 501 and a portion 502 deflected in said plane vertical 500 represented in projection.

The so-called vertical part 501 of the well 510 symbolizes a depth interval for which the value of the vertical inclination is below the effective operating threshold value of a double gimbals device, or a gravity-pendulum measuring system. the so-called deflected portion 502, corresponds to an interval for which the vertical deflection angle of the well is above said threshold value, and contains the triadrons 521 to 523.

FIG. 5a shows the attitude of the trihedrons corresponding to a fixed mounting of the three orthogonal seismic sensors in the well probe, of which 511 and 521, where the axis of the sensor generally called Z-tool is aligned with the axis of the well and points upward: the trihedron 511 in the vertical part 501 and the trihedron 521 in the deflected part 502 thus illustrate that the angle of "relative bearing" which points the Direction of the orthogonal sensors at the axis of the well relative to the azimuth of the vertical plane 500 is random from one depth station of PSV to another. The measured value of the "relative bearing" can be exploited for the orientation of the components only in the deflected part 502 (see FIG. 4a and associated explanations).

FIG. 5c represents the attitude of the trihedrons corresponding to a mounting of the three orthogonal seismic sensors on double turret-type gimbals in the well probe, of which 512 and 522, where the axis of one of the sensors is aligned with the vertical and points upwards: the trihedron 512 and the trihedrons of the adjacent dimensions in the vertical part 501 thus illustrate that the azimuthal direction of the orthogonal sensors at the axis of the well is random from one depth station of PSV to another. In contrast, the trihedron 522 in the deflected portion 502 illustrates that the orientation of the trihedron is fully known, one of the horizontal components being in the vertical plane 500 of the deviated part of the path of the well, and the other component horizontal plane being normal to plane 500 (see FIGS. 3a and 3b and associated explanations).

FIG. 5b represents the known single orientation of the trihedrons obtained after application of the orientation procedures according to one of the modes of the invention, of which 513 and 523, where the axis of one of the sensors is aligned with the vertical and pointing upwards, one of the horizontal components being in the vertical plane 500 of the deflected portion of the well trajectory and pointing in the azimuth of the increasing depths (identical to the direction of deflection of the well in the present case) , and the other horizontal component being normal to the plane 500: the trihedrons 511, 512 and the trihedrons of the adjacent random orientation dimensions in the vertical part 501 are reoriented in a common coordinate system of the triad 513 or 523 using the procedure P1 551 according to the invention. In contrast, the trihedron 521 in the deflected portion 502 is reoriented in the directions of the trihedron 523 with two successive rotations according to the known procedure P2 552 and described above (comments of Figures 4a and 4b). The trihedron 522 in the deflected portion 502 is naturally oriented identically to the triad 523 and its components require no intervention.

In practice, the procedure P1 551 according to one of the modes of the invention is also applied to the trihedra of the deflected part 502 immediately adjacent to the vertical part 501, in a short overlap and clutch interval, so as to calibrate the azimuth of the horizontal components of the trihedrons of the vertical part, including 511 and 512, on the known azimuth of the trihedrons of the deflected part 502. Finally, if it is desired to orient all the trihedrons 513 to 523 shown in the figure 5b in a geographical reference, apply an azimuthal rotation similar to that previously described (comments of Figure 3b).

Applications of the invention The method according to the invention can be applied in the framework of seismic prospection by conventional PSV method with very small offset of the single source position, in order to position in the three-dimensional space of the geological events in the vicinity of wells. Such a seismic prospecting method then comprises the following steps: the reception by triaxial seismic sensors, arranged in a well and coupled with the formations surrounding the well, in order to measure as accurately as possible the vector signal in three components of the direct waves and reflected in the P, S modes as well as the converted wave modes. the spatial orientation of multi-axis seismic receivers. To do this, the orientation method according to the invention is used. seismic well imaging from three oriented components, such as that described for example in US Pat. No. 6,076,045: this method makes use of the isotropic treatment of the three oriented components, making it possible to read the polarization of the observed reflected events, then imaging and positioning in the space of the corresponding reflectors, thereby restoring the dip and the 360-degree dip azimuth of each of the reflectors.

An important application of the method according to the invention also concerns the improvement of the quality control of the three components recorded on the recording site, using the available computer means: firstly, the computer methods, allowing three-component data orientation in a single reference frame, are easy to implement, and secondly it is easier to visually assess the overall quality of registration and overall smooth operation of the acquisition chain on oriented replay of the three components versus undirected raw replay for any depth. Thus, the orientation can be used in a single benchmark in an automated way, to obtain a quality control of the seismic measurement in three components, immediately after the acquisition of measurements in the field.

The method allows to orient the three components of PSV, in the depth intervals close to the vertical, especially when only one source position surface seismic located near the drilling rig was used, and the PSV measuring tool lowered into the well is not coupled to a precise tool for measuring all angles allowing the orientation of the three components signals in a geographical landmark. This corresponds to the usual configuration of PSVs in exploration or production wells. The method according to the invention applies effectively on a downshear shear wave train, including in the presence of propagation birefringence anisotropy: indeed, the azimuthal direction does not vary in the presence of speed anisotropy of the two own wave modes S ₁ whose effect is very small on adjacent depth levels, provided that the differential attenuation between the two waves also remains low, which is generally verified by experiment.

The method also provides guidance for the three components of PSV tools including multiple levels of simultaneous 3C seismic measurement depth, for which a single level (or an incomplete number of levels) is coupled to a full or partial orientation measurement tool. . The simplicity of implementation of the method, by means of increasingly powerful computers embedded in the acquisition systems, allows an improvement of the overall on-site quality control of the recorded three-component data, thanks to the production in slightly delayed time. , even in real time, from the point-time of the downward S-wave and from a replay of the three components oriented in a single reference frame, allowing the acquisition engineer to quickly detect on-site and with increased reliability the possible malfunctions of the chain of acquisition of the three components.

The advantage of the method is to subsequently allow the isotropic processing of the three-component PSV signals, including for old PSV data set reprocessing for which the downhole tool was not coupled to an orientation measuring tool. complete or partial.

Another advantage of the method is to allow the operator who plans to record a PSV, to refine the choice of the type of well seismic tool as well as the desired orientation tool to combine, before initiate the actual on-site acquisition of the PSV into three components, depending on the geological objective pursued, the trajectory deviation of the well considered, and the type of treatment (1C or 3C) desired following the acquisition of the field data.

The method is applicable to several geometric well seismic acquisition configurations, but specifically to vertical VSP with a weak deviation, with a source placed at a short distance from the wellhead, a configuration for which there is no known alternative to the method according to the invention. Thus, the method is applicable to very common cases where no complete and accurate orientation measuring tool is coupled to the PSV measuring tool, for example when the PSV tool comprises three components of orthogonal directional seismic sensors only. in the following configurations: a) 3C seismic sensors fixedly mounted in the PSV tool, b) 3C seismic sensors fixedly mounted in a PSV tool further comprising a device for measuring the angle of "Relative Bearing" In the plane orthogonal to the axis of the tool PSV: c) seismic sensors 3C on double gimbals with so-called "turret" architecture, that is to say comprising an axis of free rotation parallel to the axis of the tool, so parallel to the axis of the well at the anchoring station of the PSV tool. Each sensor is mounted together with a mass that is off-center with respect to the axis of the gimbals so as to obtain a pendulum device that is oriented by gravity in a known reference point related to the trajectory of the well, which is assumed to be known, for example to from a log of the measurement of the trajectory of the well by gyroscope, carried out separately.

Naturally, the type of device commonly called "Relative Bearing Sensor", as well as the mounting of seismic sensors on double gimbals mounted in "turret" are inoperative in strictly vertical wells and restore an orientation of the horizontal seismic components that is significant that beyond a low value of the order of 10 degrees of the vertical inclination of the well, and which becomes more and more precise when the inclination of the deflected well increases.

The method according to the invention can also be applied with advantage on the descending P-wave train interfered with in a vertical well, and whose form of the three-component signal varies progressively as a function of depth, but with an azimuth direction of stable total energy for the interfered signal, and in the case where an old tool having three fixed mount components in the tool, without an orientation measuring device is placed in a horizontal drain, and where the direct arrival P shows no energy on the axial component at the well. The method according to the invention can also be applied for configurations

PSV type "walkabove", when the source is located substantially in line with a horizontal drain (Figure 2b), the sensors being fixedly mounted in a well tool that has no orientation device. After maximizing the direct arrival P, we can, as a first approximation, assume that this arrival is confused with the straight line connecting the source and the receiver, from which the relative bearing angle can be deduced from knowledge of the well trajectory and the relative position of the source relative to the well.

The method according to the invention can also be applied with advantage, in order to automate the P-wave pointing and the determination of the orientation of the sensors, in the framework of conventional walkaway seismic prospecting. According to this type of method, the well receiving device may be fixed or not, and the surface source is successively activated at neighboring positions, either on a fixed azimuth line (walkaway 2D), or on a circle concentric to well or at the average geographical position of the well sensors (walkaround). The two previous configurations can be combined, either on a more or less complete grid of positions in the vicinity of the well (walkaway 3D or 3D-VSP). In particular, the method according to the invention has the advantage of providing a precise and automatic point of the direct wave P when it arrives orthogonally to the known vector direction component (substantially vertical in this case), without having to orient previously the horizontal components, in the configuration illustrated in Figure 2c.

A particular application of the method according to the invention consists in mounting a three-component PSV tool in combination with another logging tool whose orientation is desired to be known, in the extreme case or the usual orientation tools of the Like gyroscope or magnetometer / inclinometers are no longer operative, for example when the temperature of the well exceeds 220 ⁰ C.

According to a particular embodiment, for reasons of ease of calculation and reliability of the result, single markers are determined on different portions of the well by means of the method according to the invention. These landmarks have a common axis but can be of different orientation. The portions of the well have overlapping areas, which allow to determine a rotation angle to be applied to the unique markers of each portion, so as to obtain a single reference for the entire well.

Finally, the technique of pointing a filtered module signal calculated from the raw components, measured by a dipole sonic logging tool, or quadrupole in the form of a complete waveform, called "full waveform", may be useful in the case where one only wants to know the slowness and the attenuation of a shear wave without looking for the characteristics of azimuth anisotropy. In such a case, it is not useful to measure the orientation of the sonic tool in the well, which alleviates the log measurement operation.

Claims

A method of preprocessing seismic data acquired using a vertical seismic seismic survey method that includes seismic wave emission and receiving said seismic waves using at least one multi-component sensor positioned at the at least two depths, said sensor comprising at least three orthogonal geophones recording as a function of time a first seismic component in a known vector direction, and at least two other seismic components in two directions orthogonal to said known vectorial direction, characterized in that a) a new signal is constructed by calculating the square root of the sum of the squares of at least two orthogonal seismic components, referred to as the modulated signal, and time points are made the arrival of a direct seismic wave on an amplitude extremum of said module signal; b is oriented said seismic components in a single frame regardless of the depth of said sensor, by means of the following steps repeated for each depth:

a time window is defined on each side of said arrival times; an azimuthal direction is determined by maximizing an energy of said seismic components orthogonal to said known vector direction within said time window;

2. Method according to claim 1, wherein said module signal is constructed by calculating the square root of the sum of the squares of said two seismic components orthogonal to said known vector direction, and the arrival times of a wave are pointed out. downward shear.

The method of claim 2, wherein the arrival of a pressure wave is also pointed at, and velocities of said shear and pressure waves are calculated from said arrival times, and inferred speed and / or a Poisson's ratio.

4. Method according to claim 1, wherein said module signal is constructed by calculating the square root of the sum of the squares of the three seismic components, and the arrival times of a direct pressure wave are pointed out.

5. Method according to one of the preceding claims, wherein, prior to step a), the isotropy of the signal is preserved in three components, respecting amplitude ratios and phase differences between the seismic components.

6. Method according to one of the preceding claims, wherein improves a signal-to-noise ratio of the three components of the raw signal, prior to the calculation of the module, by means of an isotropic deconvolution of the three components by a single wave signal descending pressure extracted from the known vectorial direction seismic component.

7. Method according to one of the preceding claims, wherein said modulated signal is filtered so as to remove low frequency components, before pointing the arrival times of the direct seismic wave.

8. Method according to one of the preceding claims, wherein determining the geographical orientation of said single marker.

9. The method of claim 8, wherein determining the geographical orientation of said single mark by mounting said multi-component sensor on a double gimbals system, said system for directing by gravity said seismic components when the inclination of the well. reaches a value of at least about 10 degrees.

The method according to claim 8, wherein a well measurement tool including said fixedly mounted multi-component sensor is lowered into the well and the geographical orientation of said unique marker is determined by means of a system. measuring a "relative bearing" angle mounted on said measuring tool, which allows to find the orientation of said multi-component sensor when the inclination of the well reaches a value of at least about 10 degrees.

11. The method according to claim 8, wherein a well measurement tool comprising a plurality of multi-component sensors placed at adjacent measurement depths is lowered into the well, and the geographical orientation of said single coordinate system is determined by coupling to the least one of the multi-component sensors to a geographic orientation measuring tool, such as a magnetometer-inclinometer or a gyroscope.

The method of claim 8, wherein at least a portion of the well is substantially horizontal and said multi-component sensor is fixedly mounted. in a measuring tool lowered into the well, the geographical orientation of said single coordinate system is determined by assimilating a direction of maximization of a direct pressure wave to a line connecting a position of said sensor to a position of a source emitting said waves. seismic.

Method according to one of the preceding claims, wherein different markers having a common axis are determined on different portions of the well, said portions having overlapping areas for determining an angle of rotation to be applied to said single markers. in order to obtain a single reference for the entire well.

14. Method according to one of the preceding claims, wherein the orientation is used in a single frame automatically to obtain a quality control of said seismic components, immediately after acquisition of field measurements.