WO2019145746A1

WO2019145746A1 - Method for generating a reconstructed seismic signal

Info

Publication number: WO2019145746A1
Application number: PCT/IB2018/000383
Authority: WO
Inventors: Marion LECLERCQ JUILLA
Original assignee: Total Sa
Priority date: 2018-01-23
Filing date: 2018-01-23
Publication date: 2019-08-01

Abstract

A method for generating a reconstructed seismic signal from measurements recorded with detectors vertically offset from a source is presented. Seismic measurements, comprise two sets of seismic traces, each set being associated with a different position of the source. A first trace A_i(t) from a first set corresponds to an acoustic wave recorded by a first detector and a second trace Fbt_Di(t) from a second set corresponds to an acoustic wave recorded by the first detector, the first detector being located between a first and second positions of the source. The reconstructed signal corresponding to a virtual acoustic wave emitted by the source at the first position and recorded with a virtual detector located at the second position is generated from a convolution A_i(t)*Fbt_Di(t) of the first trace with the second trace.

Description

METHOD FOR GENERATING A RECONSTRUCTED SEISMIC SIGNAL

FIELD OF THE INVENTION

The invention pertains to the field of geoscientific characterization of a subsoil region. In particular, the invention addresses the issue of providing a seismic signal that can be used to determine geological parameters of a subsoil region from seismic measurements recorded in a configuration in which detectors are vertically offset from a source.

BACKGROUND OF THE INVENTION

Geological surveys generally consist in recording seismic waves emitted from at least one source that shoots acoustic waves from different positions above a subsoil region. The seismic waves are recorded by a plurality of detectors. Each detector records a seismic trace on which the time of arrival of different echoes of the acoustic wave from reflectors in the subsoil region (i.e. interfaces between layers) can be identified. These seismic traces can be processed using numerical tools to determine geological parameters of the subsoil region. Recorded seismic traces can for example be processed to provide an insight into the layer structure of the subsoil and enable the extraction of information such as, for example: acoustic wave propagation speed, rock density, composition of the subsoil region, porosity.

Existing numerical tools used to process seismic data comprise algorithms capable of building a model of the subsoil compatible with the times of arrival of different acoustic waves identified in the recorded seismic traces.

One efficient way of processing seismic signals consists in producing“common midpoint gathers”. Common midpoint gathers are sets of seismic traces corresponding to signals collected by a source and detector that are gradually offset from one another, while always being centered on the same subsoil region. These common midpoint gathers can be used to extrapolate velocities in the different identified layers of the subsoil and provide redundant data that can be “stacked” (summed) to improve the signal-to-noise ratio of the seismic data.

Existing tools have been developed to process data recorded by detectors arranged on the upper surface of the subsoil region. The source which emits acoustic waves is also located on that upper surface. In such an experimental setup, the detectors record seismic waves that arrive from the bottom of the subsoil region.

Geological surveys conducted on solid ground typically result from sources shooting seismic waves from the surface, and geophones or other detectors of acoustic waves recording the reflected waves at the same surface. Such setups do not involve significant vertical offsets between the source and the detector (if there is a small vertical offset of a few meters, it is negligible compared to the depth of the studied events). Typically, a vertical offset of about 50 to 100 meters would not have a significant impact on the processing of seismic data using conventional inversion techniques.

Another type of geological survey can be conducted to gather seismic data from a subsoil region below a seabed. These surveys may for example involve at least one source and detectors arranged on a cable or“streamer” pulled by a boat. When both source and detectors are pulled by a boat, some minor vertical offset between the source and the detector may appear (typically of the order of a few meters to a about 50 meters at most), but this has no significant effect on the nature of the seismic traces that are recorded and the signal recorded by the detectors is substantially the same as the one recorded from surveys conducted on solid ground.

Such surveys generally involve large cables or streamers that require a significant free area at the surface of the sea to safely operate the boat. As a result, it is difficult to probe subsoil that is located in the immediate vicinity of zones subject to maritime traffic or near existing platforms.

To overcome this challenge, it is possible to use cables or nodes resting on the seabed and linked to the boat pulling at least one source. These surveys are called OBC (for Ocean Bottom Cable) or OBN (for Ocean Bottom Node) and do involve a larger vertical offset between the source and the detectors. This vertical offset gives rise to some unwanted echoes in the seismic traces attributable to acoustic waves traveling from the source to the detectors without being reflected in the subsoil region. These parasite signals can be generally removed from the seismic traces because the properties of the sea medium are known or can be estimated with a reasonable degree of accuracy. Therefore, existing tools to process seismic data can still be applied to these seismic traces.

Recently, geoscientists started exploring new possibilities to conduct seismic measurements that do not require detectors to be placed on the seabed or on the surface of a subsoil region. One promising new technology consists in putting detectors in a horizontal well drilled in one of the layers of the subsoil region. Such a setup may provide more accurate data on the structure of the subsoil and may improve resolution.

However, these new setups raise new challenges that need to be addressed in order to efficiently decouple seismic waves originating from reflections below the horizontal well from those originating from above the horizontal well, in order to be able to use existing tools to process the seismic traces and obtain information about the subsoil region. Article“Metrology of a Dataset Acquired with an Optical Fiber in a Horizontal Weir by Marion Leclercq et al., Society of petroleum Engineers SPE-183170-MS provides one example of such a survey conducted using an optical fiber installed in a horizontal well. In order to process seismic data acquired in this way, only parts of the recorded signals could be interpreted and only with the help of additional data obtained using“traditional” seismic surveys to extract realistic velocities for the layers of the subsoil region.

There is therefore a need to provide an efficient method which renders seismic measurements acquired with detectors vertically offset from at least one source usable to determine a geological parameter of a subsoil region, for example to make such measurements compatible with existing tools used to process seismic data.

SUMMARY OF THE INVENTION To overcome the above-mentioned drawbacks, the invention provides a method for generating a reconstructed seismic signal usable to determine a physical parameter of a subsoil region from seismic measurements recorded with a plurality of detectors vertically offset from at least one source, the method comprising:

a\ obtaining the seismic measurements, the seismic measurements comprising at least two sets of seismic traces, each set of seismic traces being associated with a different position of the at least one source, each position of the at least one source having a horizontal coordinate along a horizontal axis, each seismic trace of each set of seismic traces corresponding to an acoustic wave recorded by a different detector from the plurality of detectors and emitted by the at least one source, the at least two sets of seismic traces comprising a first set of seismic traces associated with a first position of the at least one source having a first horizontal coordinate and a second set of seismic traces associated with a second position of the at least one source having a second horizontal coordinate,

b\ selecting a first seismic trace A,(t) from the first set of seismic traces corresponding to an acoustic wave recorded by a first detector from the plurality of detectors and a second seismic trace Fbt_Di(t) from the second set of seismic traces corresponding to an acoustic wave recorded by the first detector, the first detector being at a first detector position having a first detector horizontal coordinate between the first and second horizontal coordinates, and

c\ generating the reconstructed seismic signal corresponding to a virtual acoustic wave emitted by the at least one source at the first position and recorded with a virtual detector located at the second position by computing a convolution Ai(t) * Fbtoi(t) of the first seismic trace with the second seismic trace.

The invention provides a method for reconstructing a seismic trace which compensates the effects of the vertical offset between source and detector (and in particular in the case of large vertical offsets typically above 50 meters or 100 meters), so that the reconstructed seismic trace can be used to generate a common midpoint gather, or extract any relevant geophysical information in the subsoil region.

The term “at least one source” implies that the seismic measurements can be conducted by moving one source from a first position to several other positions, or else that the measurements can be conducted with a plurality of sources each placed at different positions.

The“plurality of detectors” can be made up of geophones, hydrophones, an optical fiber or any other detector capable of detecting seismic waves. In particular, it is possible to use the method described above to compensate for the offset between detectors and source in the context of OBC or OBN surveys, or in other configurations. The above described method becomes particularly advantageous when used on measurements originating from detectors placed in a horizontal well.

The terms“first horizontal coordinate” and“second horizontal coordinate” refer to the positions of the source projected on a horizontal plane or axis. Source and detectors are vertically offset, which results in different coordinates along a vertical axis. The positions of the source and detectors can be identified by referring to their horizontal coordinates along a same axis or in a same plane. In most setups, the source and detectors are located in a same vertical plane and the seismic measurements are then considered as two-dimensional (2D). These 2D measurements can be recorded in 2D planes that are close to each other and substantially parallel to each other to generate a set of several 2D measurements that provide insight into the structure and physical parameters of a three dimensional volume of the subsoil. The method of the invention transforms seismic traces obtained from experiments in which the source and the detector are vertically offset (and is more particularly applicable in situations in which the vertical offset is typically larger than about 100 meters) so that all existing tools to process seismic data can be used on the reconstructed seismic trace to extract information about the subsoil region.

The term “physical parameter” may encompass geological, geophysical, geoscientific or petrophysical parameters.

The above-described method can be used to lower the cost of four dimensional seismic monitoring of the evolution of the structure of a subsoil region, by relying on novel measurement techniques which do not require expensive detectors to be pulled by streamers at sea. Instead, the method of the invention can be implemented on data acquired by optical fibers in horizontal wells (or any other type of sensors densely sampled and located in a well) In particular, the invention may further output a physical parameter determined using the reconstructed seismic signal that is generated by the above-described method.

According to an embodiment, the method may further comprise, after generating the reconstructed seismic signal:

- processing the reconstructed seismic signal to determine the physical parameter of the subsoil region.

In particular the geological parameter may be chosen from among: an acoustic wave propagation speed in a layer of the subsoil region, a density of a layer medium of the subsoil region, a layer structure of the subsoil region, a composition of a layer of the subsoil region.

According to an embodiment, the seismic measurements may be recorded with the plurality of detectors placed in a horizontal well in the subsoil region.

According to an embodiment, the plurality of detectors may be within an optical fiber, the seismic measurements being recorded using distributed acoustic sensing.

Optical fibers used as detectors of seismic waves using distributed acoustic sensing offer the potential of sensing smaller details of the subsoil region when compared to standard geophone equipment owing to a very high resolution. Indeed, every portion of the optical fiber is sensitive to P-waves which makes an optical fiber a virtually continuous arrangement of detectors. It is possible to stack portions of the recorded signal in order to recreate a sampled signal and to increase the signal-to-noise ratio. One possible embodiment consists in partitioning the optical cable in contiguous one-meter long portions, and adding up the signal recorded in every portion. This recreates a sampled signal similar to that recorded when two neighboring detectors are separated by one meter. Other sampling lengths can be selected at will. The smaller the sampling length the better the imaging. It should further be noted that instead of using an optical fiber, any type of sensor that is densely sampled and placed in the well could be used.

According to an embodiment, the method may further comprise:

d\ repeating b\ and c\ iteratively by replacing, at each iteration, the first detector with a current detector, the current detector being at a current detector position having a current detector horizontal coordinate between the first and second horizontal coordinates in order to obtain a set of reconstructed seismic signals forming a reconstructed seismic trace S_ADW at a common midpoint associated with an offset between first and second positions.

The method described above provides a signal which gives insight into the presence or absence of reflectors at a given depth along a common midpoint axis of the first and second positions. The common midpoint axis is the perpendicular bisector of the segment joining the first and second positions of the at least one source. To get a broader picture of the presence or absence of reflectors in the subsoil region along the common midpoint axis, the above method is implemented on the signal recorded by each detector situated between the first horizontal coordinate and the second horizontal coordinate, and in particular all detectors situated between the first horizontal coordinate and the common midpoint axis.

According to an embodiment, the seismic measurements comprising more than two sets of seismic traces, the method may further comprise:

e\ repeating b\ and c\ iteratively to create a common midpoint gather by replacing, at each iteration, the first set of seismic traces and the second set of seismic traces with two different sets of seismic traces from the seismic measurements,

the two different sets of seismic traces being associated respectively with a current third position of the at least one source having a current third horizontal coordinate and a current fourth position of the at least one source having a current fourth horizontal coordinate, the current third position and the current fourth position having a same common midpoint axis as the first and second positions, each iteration providing a reconstructed seismic trace S_AO’^) at a common midpoint associated with an offset between the current third position and the current fourth position.

To create a common midpoint gather the method of the invention is repeated for pairs of positions of the at least one source centered on a same common midpoint axis.

According to an embodiment, the method may further comprise:

- estimating a range of times of arrival for signals reflected in the subsoil region that can be identified in one among the reconstructed seismic signal and a seismic trace from the at least two sets of seismic traces based on a position of the first detector having a first detector horizontal coordinate and a distance between said first detector horizontal coordinate and a horizontal coordinate of a common midpoint axis of the first and second positions,

- removing from the first seismic trace A,(t) and the second seismic trace Fbt_Di(t) portions of signals that can be associated with times of arrival outside of the estimated range.

Reconstructed seismic traces using the method described above comprise some echoes from waves arriving from above the detectors. These echoes overlap with the echoes of waves reflected below the position of the detectors and add noise to the reconstructed data. In order to remove these echoes, it is possible to select a range of velocities that is realistic for a given layer profile and disregard all echoes corresponding to propagation times that fall outside that range. Doing so from uppermost layers downwards for each subsequent echo in the reconstructed seismic trace can remove some of the unwanted echoes resulting from reflections happening above the detectors.

According to an embodiment, the range of times of arrival is identified using a predicted propagation time for acoustic waves emitted from the at least one source using an estimated velocity model. This estimated velocity model can for example be obtained from previous measurements recorded using more conventional setups, or from drills or other geological and/or geophysical data available about the subsoil region.

According to an embodiment, a velocity model can be estimated from a slope of the common midpoint gather.

The slope of the asymptote of the time of arrival of echoes provides the inverse of the velocity in the layer in which the reflection of the seismic wave occurred. With reconstructed seismic traces, multiple echoes from unwanted reflections in the layers above the detectors may render the identification of the time of arrival of echoes more complicated, as these parasitic echoes may partially overlap with the relevant echoes. However, even if the data is noisy, a slope can be estimated within a reasonable error range from a common midpoint gather.

According to an embodiment, the second seismic trace Fbt_Di(t) may be replaced with a portion of the second seismic trace comprising only an acoustic wave signal reconstructed using a time of first arrival recorded at the first detector.

The convolution which leads to a reconstructed seismic trace described above does not require the whole seismic signal collected by the first detector and emitted from the second position. The time of the first recorded wave, corresponding to the time of the first recorded acoustic wave which follows a direct path from the source to the detector, provides the required information to reconstruct the signal. The time of the first recorded wave is input as a Dirac signal with which the signal A(t) is convoluted.

When the entire signal of the second seismic trace is used, all the echoes recorded after this first recorded wave may scramble the seismic signal and add noise to it.

According to an embodiment, the method may further comprise:

- stacking seismic traces recorded by a predetermined number of successive detectors to improve a signal-to-noise ratio. Adding up signals recorded by different neighboring detectors is particularly advantageous in the case seismic measurements recorded using an optical fiber.

According to an embodiment, the method may further comprise:

- changing the predetermined number of successive detectors and repeating a\ through c\.

By doing so, it is possible to group detectors together in different ways, in order to improve signal-to-noise and to confirm or infirm data which seems inaccurate from previous implementations of the method of the invention.

According to an embodiment, the method may further comprise:

- selecting a third seismic trace D,(t) from the second set of seismic traces corresponding to an acoustic wave recorded by a second detector from the plurality of detectors and a fourth seismic trace Fbt_Ai(t) from the first set of seismic traces corresponding to an acoustic wave recorded by the second detector, the second detector being located at a second detector position having a second detector horizontal coordinate between the first and second horizontal coordinates, the first detector and the second detector being substantially equally spaced apart from a common midpoint axis of the first and second positions,

- determining the reconstructed seismic signal by calculating an average of the convolution A,(t)* Fbt_Di(t) of the first seismic trace with the second seismic trace and a convolution D,(t) * Fbt_Ai(t) of the third seismic trace with the fourth seismic trace.

Since acoustic waves travel along the same path in both directions, a reconstructed seismic trace corresponding to a virtual acoustic wave emitted by the at least one source in the first position and detected by a virtual detector located at the second position is equivalent to a virtual acoustic wave emitted by the at least one source in the second position and detected by a virtual detector located at the first position. It is then possible to take into account an average of both of these seismic traces to reduce contribution of noise that may be present in each reconstructed seismic trace. The term“substantially” equally spaced typically indicates that some error in the positioning of the source and detectors is admissible, for example to take into account the sampling distance in the plurality of detectors or in the distance separating the virtual detectors.

The invention is also directed to a non-transitory computer readable storage medium, having stored thereon a computer program comprising program instructions, the computer program being loadable into a data-processing unit and adapted to cause the data-processing unit to carry out the steps of the above- described method when the computer program is run by the data-processing device.

BRIEF DESCRIPTION OF THE DRAWINGS

The method of the invention will be better understood by reading the detailed description of exemplary embodiments presented below. These embodiments are illustrative and by no means limitative. They are provided with the appended drawings in which:

- figure 1 is a schematic representation of a subsoil comprising several layers with a plurality of detectors in the subsoil and a boat pulling at least one source close to the sea surface;

- figure 2a is a schematic representation in two-dimensions of the setup of figure 1 , with illustrations of rays emitted from two sources and recorded by detectors on an optical fiber;

- figure 2b is a diagram showing reconstructed signals obtained using traces recorded by the detectors of figure 2a as well as a reconstructed seismic trace that can be obtained using these signals;

- figure 3a is a schematic representation in two-dimensions of the setup of figure 1 , with illustrations of rays emitted from two sources and recorded by detectors in an optical fiber similar to that of figure 2a, but with a different offset between the two sources and so that the two sources of figure 3a have the same common midpoint axis as the two sources of figure 2a; - figure 3b is a diagram showing reconstructed signals obtained using traces recorded by the detectors of figures 2a and 3a as well as the reconstructed seismic traces that can be obtained using these signals;

- figure 4a is a schematic representation in two-dimensions of the setup of figure 1 , showing unwanted reflections above the detectors;

- figure 4b is a schematic representation of a reconstructed seismic trace comprising unwanted reflections from above the detectors from figure 4a;

- figure 5a is a schematic common midpoint gather comprising reconstructed seismic traces with overlapping unwanted echoes from above the detectors;

- figure 5b is a schematic common midpoint gather comprising improved reconstructed seismic traces from which unwanted echoes from above the detectors were removed;

- figure 6 is a schematic representation of the trajectory of seismic waves reflected in the subsoil region from different depths of reflectors, showing that the distance between detectors from the plurality of detectors in an optical fiber affect the depth of the subsoil region that can be probed;

- figure 7 is a possible embodiment for a device that enables the method of the invention to be implemented.

For the sake of clarity, the dimensions of features represented on these figures may not necessarily correspond to the real-size proportions of the corresponding elements. Like reference numerals on the figures correspond to similar elements or items.

DETAILED DESCRIPTION OF THE INVENTION

The invention provides a novel means of processing seismic data to use the full potential of promising technologies providing new and efficient ways of probing a subsoil region. In particular, the invention provides a method which reconstructs seismic signals so that they can be processed using existing tools to output a physical parameter from seismic measurements acquired with detectors vertically offset from sources.

Figure 1 provides one illustration of a geological survey in which at least one source 101 is pulled by a boat 10 substantially at the surface 100 of a sea medium 111. The at least one source 101 can be one source that is moved by the boat to occupy different positions 102-105 or several sources moved by the boat and/or arranged on a streamer and activated in sequence. The at least one source 101 of figure 1 is moved to occupy different positions 102-105 indicated in the two- dimensional representation of figure 1 by horizontal coordinates 11 -15 along a horizontal axis 1 or x axis. The at least one source 101 sends acoustic waves 1001 -1005 into the sea medium which penetrate different layers 121 -125 of a subsoil region 112. When travelling through these different layers the acoustic waves 1001 -1005 can be reflected at interfaces 110, 120, 130, 140, 150 and be recorded by a detector 201 that is part of a plurality of closely packed detectors. In figure 1 , the plurality of detectors is represented by an optical fiber 200 inserted in a horizontal well 999. However, any other types of detectors might be used. It is further to be noted that the horizontal well may follow a substantially horizontal path in the subsoil region, but comprise minor lateral or vertical offsets.

One advantage of using an optical fiber 200 is that it provides a very high resolution, since every portion of the optical fiber is sensitive to compressive P acoustic waves. The optical fiber 200 can be regarded as a continuous detector arranged in a horizontal well. Distributed acoustic sensing is a technique that allows recording a signal with an optical fiber 200, capable of associating a recorded signal with a position on the fiber that recorded it.

An optical fiber 200 can be partitioned into portions, to act as a set of individual detectors, two neighboring detectors being spaced apart by a sampling distance that can be selected by choosing the length of each portion. The signals recorded in each portion are summed up or“stacked” to produce a signal with a higher signal-to-noise ratio. A typical sampling distance can be set to 1 meter.

As seen on figure 1 , the at least one source 101 is vertically offset from the plurality of detectors. A vertical axis 2 or z axis shows the offset between sources and detectors. The at least one source 101 is at a first depth 20 or zO whereas the plurality of detectors are substantially located at a second depth 21 or z1.

It is to be noted that although figure 1 represents a geological survey conducted on a subsea subsoil region 112, the same type of survey could also be conducted on a subsoil region below solid ground exposed to air.

As can be seen on figure 1 , positioning the plurality of detectors within the subsoil region 112 gives rise to the recording of some unwanted acoustic waves 1002. These unwanted acoustic waves 1002 are echoes of waves occurring above the detector 201.

During a geological survey, seismic measurements are recorded by emitting an acoustic wave from each of the different positions 102-105. For each of these positions, all the detectors from the plurality of detectors detect the emitted seismic waves as well as its echoes from interfaces (or reflectors) in the subsoil region 112. Each detector records a seismic trace on which the times of arrival of the seismic waves and its echoes appear. For each position of the at least one source, a set of seismic traces is thus obtained.

The purpose of the invention is to render a seismic signal recorded by the detectors of a plurality of detectors that are significantly vertically offset from the sources, such as in the survey represented on figure 1 , usable to determine a geological (or physical) parameter of a subsoil region 112 therefrom. The term “significantly vertically offset” refers to offsets of more than about 50 or 100 meters. This involves correcting the effect of the vertical offset between the sources and detectors, and may in particular involve correcting the signal recorded by each detector of the plurality of detectors and optionally to identify echoes originating from above the detectors and echoes originating from below the detectors.

Figure 2a illustrates a simplified version of the experimental setup of figure 1. In this figure, the at least one source emits seismic waves from a first position 107 and later from a second position 104. Six positions 103-108 are represented on figure 2a. The first position 107 has a horizontal coordinate 17 and a vertical coordinate 20. The second position 104 has a horizontal coordinate 14 and the same vertical coordinate as the first position. In fact, all positions of the at least one source advantageously have substantially the same vertical coordinate.

The invention assumes that the seismic wave that is generated by the at least one source from each position 103-108 is substantially the same.

A purpose of the method of the invention is to generate a reconstructed seismic signal that corresponds to the one emitted by the at least one source from the first position 107 and that would have been detected by a detector placed at the second position 104.

To do that, a first seismic trace Aj(t), corresponding to the acoustic wave emitted by the at least one source in the first position 107 is selected from a first set of seismic traces. This first seismic trace A(t) is recorded by a first detector 202 which is at a position called “first detector position” having a first detector horizontal coordinate 142 that is between the first horizontal coordinate 17 and second horizontal coordinate 14.

A second seismic trace Fbt_Di(t), corresponding to the acoustic wave emitted by the at least one source in the second position 104 is selected from a second set of seismic traces. This second seismic trace Fbt_Di(t) is also recorded by the first detector 202.

Figure 2a illustrates rays of the acoustic waves emitted from the first and second positions by the at least one source, until they reach the first detector 202. A first ray 172 emitted from the first position 107 reaches a point 220 in the subsoil region at which it is reflected before being measured by the first detector 202. Other rays corresponding for example to the direct path between the first position 107 and the detector 202 without being reflected by the subsoil region are also recorded on the first seismic trace A(t). Flowever, the first ray 172 provides some insight into the structure of the subsoil at depth 22 marked z2 on figure 2a.

The second ray 1420 corresponds to the direct path between the second position 104 and the first detector 202. It can be seen that the second ray 1420 corresponds substantially to the remainder of the itinerary of the first ray 172 from the first detector 202 to the second position 104, but in reverse. To generate a reconstructed seismic signal corresponding to a ray following a path from the first position 107 to the second position 104, the method of the invention involves calculating a convolution A(t)* Fbt_Dj(t) between the first seismic trace A(t) and the second seismic trace Fbt_Dj(t).

Instead of taking all of the second seismic trace Fbt_Dj(t) recorded by the first detector 202, it is possible to take only the portion corresponding to the first arrival recorded by the first detector 202. Reducing the second seismic trace Fbt_Di(t) to the time of this first arrival reduces noise that is introduced in the convolution by all the echoes recorded by the first detector 202 of the seismic wave emitted from the second position 104. When the time of this first arrival is used instead of the second seismic trace Fbt_Di(t), the second seismic trace Fbt_Di(t) is replaced in the convolution described above by a Dirac function reconstructing an acoustic wave in the form of an impulse at the time of the first arrival.

The above convolution generates a reconstructed seismic signal that probes a small portion of the subsoil region. Indeed, it can for example provide information on whether there is a reflector at point 220 (or within a limited area around point 220) in the subsoil region. In case an interface between two layers crosses point 220, the reconstructed seismic signal shall have the signature of an echo of the acoustic wave emitted from the first position 107.

To probe more points in the subsoil region, the first detector 202 is replaced by any other detector located between two extremes corresponding to detectors 204 and 207 on figure 2a. Each detector having a horizontal coordinate lying between the first horizontal coordinate 17 and the second horizontal coordinate 14 can be used to probe a different depth of the subsoil region along a common midpoint axis 4500 of the first and second positions. The common midpoint axis 4500 corresponds to the perpendicular bisector of a segment joining the first position 107 with the second position 104.

Figure 2a shows a second example of a ray 173 corresponding to the seismic acoustic wave emitted from the first position 107 and detected by another detector called“current detector” 203. The term“current detector” is used to designate any detector other than the first detector within the meaning of this description. The current detector 203 has a current detector horizontal coordinate 143. Another ray 1430 corresponding to the seismic acoustic wave emitted from the second position 104 and detected by the current detector 203 is also represented on figure 2a.

These two other rays correspond to two different seismic traces recorded by the current detector 203, which provides insight into the structure of the subsoil at a point 230 located at a depth 23 marked z3 on figure 2a.

Figure 2a thus illustrates that it is possible to probe different depths of the subsoil region by changing the“current detector” that is selected to find a pair of seismic traces recorded at the“current detector” and emitted respectively from the first position 107 and the second position 104. The “current detector” can be any detector having a current detector horizontal coordinate between the first horizontal coordinate 17 and the second horizontal coordinate 14.

It should be noted that to probe the subsoil region at different depths along the common midpoint axis 4500, the position of the “current detectors” can be restricted to positions having a current detector horizontal coordinate 143 between the second horizontal coordinate 14 and a horizontal coordinate 4501 of the common midpoint axis 4500. In particular, in the example represented on figure 2a in which the rays reflected in the subsoil region originate from the first position 107, the“current detector” should not be selected within a portion 1700 of the optical fiber 200, at positions having a current detector horizontal coordinate that is between the first horizontal coordinate 17 and the horizontal coordinate 4501 of the common midpoint axis 4500.

However, it should also be noted that the rays 172, 1420 represented on figure 2a are virtually equivalent to rays (not represented) that would be emitted from the second position 104 and reflected at point 220 to then be detected by a second detector 2020 and rays emitted from the first position 107 and detected by the second detector 2020. This second detector 2020 and the first detector 202 are equally spaced apart from the common midpoint axis 4500. The second detector has a second detector horizontal coordinate 1422. This other set of rays would give rise to a third seismic trace Dj(t) emitted from the second position 104 and recorded at the second detector 2020 and a fourth seismic trace Fbt_Aj(t) emitted from the first position 107 and recorded at the second detector 2020. Similarly, the rays 173, 1430 are virtually equivalent to rays (not represented) emitted from the second position 104 and reflected at point 230 to then be detected by a second“current detector” 2030, having a second detector horizontal coordinate 1433. The“current detector” 203 and the second “current detector” 2030 are equally spaced apart from the common midpoint axis 4500. The second detector 2020 and second“current detector” 2030 are within a portion 1700 of the optical fiber 200 of figure 2a.

Figure 2a represents two examples of seismic traces that can be used to generate a reconstructed seismic signal for two positions on a common midpoint axis 4500, for an offset 1470 between a first position 107 and a second position 104 of the at least one source.

The distance 211 separating two neighboring detectors can typically be set at 1 meter, even though this distance can be adjusted at will, to find an appropriate balance between improvement of the signal-to-noise ratio and the resolution.

It is for example possible to group detectors together in different ways, to add up their recorded signals in order to increase the signal-to noise ratio at the expanse of vertical resolution along the common midpoint axis 4500.

The reconstructed seismic signals obtained by the method described above are shown on a graph in figure 2b, along with a reconstructed seismic trace S_ADW 410 obtained when all the reconstructed seismic signals are added up.

Figure 2b represents the reconstructed seismic signal 312 obtained by the convoluting seismic traces corresponding to rays 1420 and 172 of figure 2a (expressed by the convolution A,(t)* Fbt_D,(t) ) and the reconstructed seismic signal 313 obtained by convoluting seismic traces corresponding to rays 1430 and 173 of figure 2a. The reconstructed seismic signals appear schematically on the left side of curve 300 in figure 2b for all the reconstructed seismic signals obtained using seismic traces recorded by“current detectors” 203 situated in portion 1400 of the optical fiber 200. Curve 300 has symmetrical reconstructed seismic signals (for example reconstructed seismic signal 3121 obtained by the convolution Di(t)* Fbt_Aj(t) and reconstructed seismic signal 3131 ) on the right side of curve 300 for all the reconstructed seismic signals obtained using seismic traces recorded by “current detectors” situated in portion 1700 of the optical fiber 200.

The graph of figure 2b displays the time of arrival of the echoes along a time axis 4 and the position of the detector along a position axis 3. This graph shows the equivalence in terms of time of arrivals of reconstructed seismic signals 312 and 3121 , or 313 and 3131.

The reconstructed seismic trace S_AD(t) 410 of figure 2b can be obtained by adding up all the reconstructed seismic signals obtained using the“current detectors” from the portion 1400 of figure 2a, or only those from portion 1700. However, it is also possible to calculate an average of the reconstructed seismic signals obtained using“current detectors” from both portions 1400 and 1700. One possibility for such an average could be expressed with the following formula:

where i designates all“current detectors” from portion 1400 in the left hand side of the above equation and j designates all“current detectors” from portion 1700 in the right hand side of the above equation.

When all the reconstructed seismic signals obtained for a given offset 1470 between the first position 107 and the second position 104 are added up, the reconstructed seismic trace S_AD(t) 410 of figure 2b is obtained. S_AD(t) 410 is a reconstructed seismic trace at a common midpoint associated with offset 1470. This reconstructed seismic trace 410 comprises echoes 412, 413, 414 indicating positions of reflectors in the subsoil region along the common midpoint axis 4500.

Seismic measurements can be further processed to generate a common midpoint gather by reproducing the method described above with another offset between two different positions of the at least one source, so that these two different positions have the same common midpoint axis 4500 as the one described above.

Such a configuration is represented on figure 3a. Figure 3a shows several positions 102-109 for the at least one source and two sets of rays reflected from points 220 and 230, emitted from a current third position 109 of the at least one source having a current third horizontal coordinate 19 and a current fourth position 102 of the at least one source having a current fourth horizontal coordinate 12. A ray 194 is emitted from the current third position 109 and recorded by detector 204. Another ray 1240 is emitted from the current fourth position 102 and recorded by the detector 204. Detector 204 is at a distance 2004 from the common midpoint axis 4500.

Another ray 195 is emitted from the current third position 109 and recorded by a detector 205, while ray 1250 is emitted from the current fourth position 102 and recorded by the detector 205.

Just as explained above in connection with figures 2a and 2b, the seismic traces recorded by detectors 204 and 205 are equivalent to those detected in similar conditions at the second detectors 2040 and 2050.

In figure 3a, the offset 1290 between the current third position 109 and the current fourth position 102 differs from the offset 1470 of figure 2a. Repeating the above method for different offsets provides data that can be used to create a common mid point gather such as those represented on figures 5a and 5b.

Figure 3b reproduces the same graph representation as that of figure 2b. Curve 300 from figure 2b corresponds to reconstructed seismic signals obtained with the offset 1470 of figure 2a. Curve 310 of figure 3b corresponds to reconstructed seismic signals obtained using seismic traces recorded by detectors from portion 1200 of figure 3a and portion 1900 of figure 3a. In particular, curve 310 represents reconstructed seismic signal 322 obtained from the seismic traces recorded by detector 204, and reconstructed seismic signal 323 obtained from the seismic traces recorded by detector 205. Similarly curve 310 represents reconstructed seismic signal 3222 obtained from the seismic traces recorded by detector 2040, and reconstructed seismic signal 323 obtained from the seismic traces recorded by detector 2050.

It can be seen that curve 310 is made up of reconstructed seismic signals that have times of arrival set later along the time axis 4 than those of curve 300. This is due to the fact that the offset 1290 of the configuration of figure 3a is larger than offset 1470 of the configuration of figure 2a. This is also reflected on the reconstructed seismic trace S_AO W 420 of figure 3b, comprising echoes 422, 423, 424 that appear later than echoes 412-414 of reconstructed seismic trace S_AD(t) 410.

By repeating the above method for different offsets, several reconstructed seismic traces at the common midpoint axis 4500 can be obtained, to generate the common midpoint gathers of figures 5a and 5b.

Figures 4a and 4b show the impact of unwanted echoes on the reconstructed seismic traces. In figure 4a, three rays are shown. Ray 186 is emitted from a position 108, reflected at point 240 in the subsoil region 112 and recorded by detector 206. Ray 136 is emitted from position 103 and recorded by detector 206. Detector 206 also records echoes of ray 1860 within layer 122 located above the optical fiber 200. These echoes appear on figure 4b in the reconstructed seismic trace 4000 as echoes 1861 , 1862, 1863. Although these particular echoes are shown as not overlapping with the echoes 4002, 4003, 4004 of reflections occurring below the optical fiber 200, other echoes may overlap with relevant signal and add noise to the reconstructed seismic trace 4000. All these unwanted echoes should be removed to improve accuracy of the determination of geological parameters using these reconstructed seismic signals and traces.

Figures 5a, 5b and 6, the content of which is discussed below, provide further illustrations of techniques that can be used to remove unwanted echoes from recorded and reconstructed seismic traces.

Figure 6 shows that the pair of detectors used to reconstruct a seismic signal using the method of the invention determines the depth of the reflector in the subsoil that can be probed. In particular, when a pair of detectors 203, 2030 that are close to each other are selected, the position of the reflector 601 that is probed by the direct wave that is recorded is located close to the well depth. In contrast to the depth probed by the pair of detectors 203, 2030; the direct waves recorded by the pair of detectors 206, 2060 spaced further apart from each other gives access to the characteristics of reflector 602, located deeper in the subsoil region than reflector 601. As a result, any echo in the reconstructed seismic signal that has a time of first arrival that is incompatible with the depth of reflector 601 , 602 that can be probed by the pair of detectors used to reconstruct the seismic signal can be considered as being an unwanted echo (or noise) and removed from the reconstructed seismic signal.

In other words, each pair of detectors used to produce the reconstructed seismic signal sets a substantially narrow time frame within which the time of first arrival of direct waves is recorded, the remaining portion of the reconstructed signal corresponding to echoes that can be removed prior to processing the reconstructed seismic signal.

A common midpoint gather 5001 for a set of reconstructed seismic traces 410, 420, 430, 440, 450, 460, 470 is shown on figure 5a. The reconstructed seismic traces are displayed as a function of the offset between the two positions of the source (axis 5). It can be seen that the effect of unwanted echoes overlapping with echoes originating from below the plurality of detectors is mostly seen for bigger offsets. This can be explained by the longer propagation time between source and detector as the offset is increased, which provides more opportunities for repeated echoes to add noise to the seismic traces.

Each reconstructed seismic trace of figure 5a is shown with a first echo 402 and a second echo 403. As can be seen, the first echo becomes blurred and polluted by noise as the offset is increased. This renders the estimation of the wave propagation speed in the layer in which the echo occurred more difficult. A typical and well-known method for estimating wave propagation speed from a common midpoint gather consists in estimating the slope of the asymptote of the arrival of each echo which corresponds to the inverse of the wave propagation speed.

In the common midpoint gather of figure 5a, such a determination can only be made within the error range 5000 that is delimited between two curves 501 and 502 around the signals corresponding to the first echo that can be identified on each reconstructed seismic trace.

An original solution proposed by the invention to remove unwanted reflections from the reconstructed seismic traces consists in removing echoes that can be identified as noise based on the times of arrival of echoes. For example, on the reconstructed seismic traces of figure 5a, signal 503 appears outside of error range 5000 and cannot reasonably be identified as corresponding to first echo 402. Therefore, this signal can be removed from the reconstructed seismic trace.

A similar approach can be implemented on the recorded seismic traces prior to their use to generate a reconstructed signal.

One possibility to identify unwanted echoes on the recorded seismic traces or on the reconstructed seismic traces is to use an available or estimated velocity model. An estimation of the acoustic wave propagation speed in the layers may be possible using available complementary surveys or based on former approximate knowledge of the structure of the subsoil region.

Using a first estimate, the recorded seismic traces can be processed to remove echoes that can be identified as corresponding to unrealistic times of arrival. The processed seismic signals can then be used to generate a reconstructed seismic signal and traces, usable to plot a common midpoint gather on the basis of which a more accurate estimation of the velocity in each layer is possible.

Even if no estimation of the velocity is available, it is possible to extract a first value for the propagation speed in each layer using a first set of reconstructed seismic signals on a raw common midpoint gather such as that illustrated on figure 5a. Several slopes can be fitted into error range 5000. One possibility is represented by slope 510, which can be used to provide a first estimate of the velocity in the first layer.

Based on this estimation, the recorded seismic traces can be processed to remove unwanted portions of signal around the echoes corresponding to a reflection in the first layer below the plurality of detectors. Doing so recursively, and in sequence for each layer enables removing noise from the seismic traces and produces a better common midpoint gather 5002 such as that of figure 5b. In figure 5b, the identification of slopes fitting echoes of improved reconstructed seismic traces 41000, 42000, 43000, 44000, 45000, 46000, 47000 is easier than on the traces of figure 5a. Based on the improved common midpoint gather 5002 of figure 5b, the estimation of the acoustic wave propagation speed in the layers of the subsoil region is more accurate, as can be seen by straight lines 511 and 512 which are not chosen with an error range as in figure 5a.

Figure 7 is a possible embodiment for a device that can be used to implement the above method.

In this embodiment, the device 700 comprises a computer, this computer comprising a memory 705 to store program instructions loadable into a circuit and adapted to cause circuit 704 to carry out the steps of the present invention when the program instructions are run by the circuit 704.

The memory 705 may also store data and useful information for carrying the steps of the present invention as described above.

The circuit 704 may be for instance:

- a processor or a processing unit adapted to interpret instructions in a computer language, the processor or the processing unit may comprise, may be associated with or be attached to a memory comprising the instructions, or

- the association of a processor / processing unit and a memory, the processor or the processing unit adapted to interpret instructions in a computer language, the memory comprising said instructions, or

- an electronic card wherein the steps of the invention are described within silicon, or

- a programmable electronic chip such as a FPGA chip (for « Field- Programmable Gate Array »).

This computer comprises an input interface 703 for the reception of data used for the above method according to the invention and an output interface 706 for providing a stacked model.

To ease the interaction with the computer, a screen 701 and a keyboard 702 may be provided and connected to the computer circuit 704. The above examples and embodiments show how the position of reflectors in the subsoil region or an acoustic wave propagation speed can be determined. However, it should be noted that one advantage of the method described above is that it recreates a seismic signal that is similar to that acquired with no or unsignificant (lower than about 50 meters or 100 meters) vertical offset between the detector and source. Therefore, once a reconstructed seismic signal or a reconstructed seismic trace is generated, available numerical tools to process seismic data can be used to extract geological parameters about the subsoil region.

Although the above description is provided in relation with figures illustrating substantially horizontal reflectors in the subsoil region, it should be noted that existing models and numerical tools to process seismic signals recorded from dipped reflectors can easily be applied to the method of the invention.

Claims

1. A method for generating a reconstructed seismic signal (312, 313, 3121 , 3131 ) usable to determine a physical parameter of a subsoil region (112) from seismic measurements recorded with a plurality of detectors vertically offset from at least one source (101 ), the method comprising:

a\ obtaining the seismic measurements, the seismic measurements comprising at least two sets of seismic traces, each set of seismic traces being associated with a different position of the at least one source, each position of the at least one source having a horizontal coordinate along a horizontal axis (1 ), each seismic trace of each set of the seismic traces corresponding to an acoustic wave recorded by a different detector from the plurality of detectors and emitted by the at least one source, the at least two sets of seismic traces comprising a first set of seismic traces associated with a first position (107, 109) of the at least one source having a first horizontal coordinate (12, 14) and a second set of seismic traces associated with a second position (104, 102) of the at least one source having a second horizontal coordinate (17, 19),

b\ selecting a first seismic trace A,(t) from the first set of seismic traces corresponding to an acoustic wave recorded by a first detector (202, 204) from the plurality of detectors and a second seismic trace Fbt_Di(t) from the second set of seismic traces corresponding to an acoustic wave recorded by the first detector, the first detector being at a first detector position having a first detector horizontal coordinate between the first and second horizontal coordinates, and

c\ generating the reconstructed seismic signal (312, 313, 3121 , 3131 ) corresponding to a virtual acoustic wave emitted by the at least one source at the first position and recorded with a virtual detector located at the second position by computing a convolution A,(t)* Fbt_Di(t) of the first seismic trace with the second seismic trace.

2. The method of claim 1 , further comprising, after generating the reconstructed seismic signal:

3. The method of any one of claims 1 or 2, wherein the geological parameter is chosen from among: an acoustic wave propagation speed in a layer (121 -124) of the subsoil region, a density of a layer medium of the subsoil region, a layer structure of the subsoil region, a composition of a layer of the subsoil region.

4. The method of any one of claims 1 to 3, wherein the seismic measurements are recorded with the plurality of detectors placed in a horizontal well in the subsoil region.

5. The method of any one of claims 1 to 4, wherein the plurality of detectors is within an optical fiber (200), the seismic measurements being recorded using distributed acoustic sensing.

6. The method of any one of claims 1 to 5, further comprising:

d\ repeating b\ and c\ iteratively by replacing, at each iteration, the first detector (202, 204) with a current detector (203, 205), the current detector being at a current detector position having a current detector horizontal coordinate between the first and second horizontal coordinates in order to obtain a set of reconstructed seismic signals forming a reconstructed seismic trace S_ADW (410) at a common midpoint associated with an offset (1470) between the first and second positions.

7. The method of claim 6, wherein the seismic measurements comprise more than two sets of seismic traces, the method further comprising:

e\ repeating b\ and c\ iteratively to create a common midpoint gather (5001 - 5002) by replacing, at each iteration, the first set of seismic traces and the second set of seismic traces with two different sets of seismic traces from the seismic measurements,

the two different sets of seismic traces being associated respectively with a current third position (109) of the at least one source having a current third horizontal coordinate (19) and a current fourth position (102) of the at least one source having a current fourth horizontal coordinate (12), the current third position and the current fourth position having a same common midpoint axis (4500) as the first and second positions, each iteration providing a reconstructed seismic trace S_AOW (420) at a common midpoint associated with an offset (1290) between the current third position and the current fourth position.

8. The method of any one of claims 1 to 7, further comprising:

- estimating a range of times of arrival for signals reflected in the subsoil region that can be identified in one among the reconstructed seismic signal and a seismic trace from the at least two sets of seismic traces based on a position of the first detector having a first detector horizontal coordinate and a distance (2004) between said first detector horizontal coordinate and a horizontal coordinate of a common midpoint axis (4500) of the first and second positions,

9. The method of claim 8, wherein the range of times of arrival is identified using a predicted propagation time for acoustic waves emitted from the at least one source using an estimated velocity model.

10. The method of any one of claims 8 to 9, dependent on claim 7, wherein a velocity model is estimated from a slope (510) of the common midpoint gather.

11. The method of any one of claims 1 to 10, wherein the second seismic trace Fbtoi(t) is replaced with a portion of the second seismic trace comprising only an acoustic wave signal reconstructed using a time of first arrival recorded at the first detector.

12. The method of any one of claims 1 to 11 , further comprising:

- stacking seismic traces recorded by a predetermined number of successive detectors to improve a signal-to-noise ratio.

13. The method of any one of claims 1 to 12, further comprising:

14. The method of any one of claims 1 to 13, further comprising:

- selecting a third seismic trace D,(t) from the second set of seismic traces corresponding to an acoustic wave recorded by a second detector (2020, 2030, 2040, 2050) from the plurality of detectors and a fourth seismic trace Fbt_Ai(t) from the first set of seismic traces corresponding to an acoustic wave recorded by the second detector, the second detector being located at a second detector position having a second detector horizontal coordinate between the first and second horizontal coordinates, the first detector and the second detector being substantially equally spaced apart from a common midpoint axis (4000) of the first and second positions,

- determining the reconstructed seismic signal by calculating an average of the convolution A,(t)* Fbt_Di(t) of the first seismic trace with the second seismic trace and a convolution D,(t)* Fbt_Ai(t) of the third seismic trace with the fourth seismic trace.

15. A non-transitory computer readable storage medium, having stored thereon a computer program comprising program instructions, the computer program being loadable into a data-processing unit and adapted to cause the data-processing unit to carry out the steps of any of claims 1 to 14 when the computer program is run by the data-processing device.