METHOD AND DEVICE FOR MONITORING THE SUBSOIL OF THE EARTH UNDER A TARGET ZONE

Abstract

In order to monitor the subsoil of the earth under a target zone, seismic waves coming from an identified mobile noise source are recorded by means of at least one pair of sensors disposed on either side of the target zone, time periods are selected corresponding to the alignments of the pairs of sensors with the noise source, a seismogram of the target zone is reconstructed by interferometry based on the recorded seismic waves and on the selected time periods and an image of the subsoil of the target zone is generated using the seismogram.

Claims

1. A method for monitoring the subsoil of the earth under a target zone, characterized in that it comprises steps according to which: seismic waves coming from an identified and mobile source of seismic noise are recorded by means of at least one pair of sensors disposed on either side of said target zone; time periods are selected corresponding to the alignments of the pairs of sensors with said noise source; a seismogram of the target zone is reconstructed by interferometry based on said recorded seismic waves and the selected time periods; and an image of the subsoil under the target zone is generated using the seismogram.

2. The method as claimed in claim 1, characterized in that the source is moving outside of the segment formed by the disposition of said pair of sensors.

3. The method as claimed in claim 1, characterized in that it furthermore comprises, prior to the reconstruction step, a step for pre-processing the recorded seismic waves.

4. The method as claimed in claim 3, characterized in that the pre-processing step comprises denoising and frequency filtering operations.

5. The method as claimed in claim 3, characterized in that the pre-processing step comprises a spectral whitening operation.

6. The method as claimed in claim 1, characterized in that, during the step for reconstruction by interferometry, at least one operation is carried out for intercorrelation of signals coming from the seismic waves respectively recorded by the two sensors of said pair of sensors.

7. The method as claimed in claim 1, characterized in that the recording step is carried out continuously over time.

8. The method as claimed in claim 1, characterized in that it furthermore comprises, following the reconstruction step, a step for addition of reconstructed signals of said seismogram coming from several mobile noise sources respectively aligned with said pair of sensors during separate periods of time.

9. The method as claimed in claim 1, characterized in that, during the recording step, seismic surface waves are recorded.

10. The method as claimed in claim 9, characterized in that, during the step for generation of said image: dispersion curves of the surface waves in the reconstructed signals of said seismogram are determined within a predetermined band of frequencies; a tomography of the speed of said reconstructed signals is carried out based on said dispersion curves; and an inversion of the surface wave speeds of said reconstructed signals is carried out, so as to produce a speed model of P- and S-waves.

11. The method as claimed in claim 1, characterized in that, during the recording step, refracted seismic waves are recorded.

12. The method as claimed in claim 11, characterized in that, during the step for generation of said image: the time of arrival of the reconstructed signals of said seismogram corresponding to said refracted seismic waves is determined; and a tomography of said reconstructed signals corresponding to the refracted P-waves is carried out based on said time of arrival, so as to produce a speed model of P-waves.

13. The method as claimed in claim 1, characterized in that, during the recording step, reflected seismic waves are recorded.

14. The method as claimed in claim 1, characterized in that said sensors comprise at least one geophone and/or at least one accelerometer and/or at least one sensor based on the use of optical fibers.

15. The method as claimed in claim 1, characterized in that said sensors disposed on either side of said target zone are placed according to at least two rows, or on square or hexagonal grids, and/or are regularly spaced.

16. The method as claimed in claim 1, characterized in that said target zone is a railroad track or a road or a building.

17. The method as claimed in claim 1, characterized in that said target zone is a first railroad track and said source is a train travelling over a second railroad track substantially parallel to the first railroad track.

18. The method as claimed in claim 17, characterized in that said sensors are disposed in a regularly spaced fashion in two parallel rows surrounding the first railroad track.

19. A device for monitoring the subsoil of the earth under a target zone, characterized in that it comprises means designed to implement steps of a method as claimed in claim 1.

Description

[0045] As shown by the flow diagram in FIG. 1, during a first step 10 of the method for monitoring the subsoil of the earth under a target zone according to the invention, seismic waves are recorded for at least a predetermined period of time.

[0046] These seismic waves originate from a mobile source of seismic noise, in other words a moving source whose motion produces noise. This noise source is known in the sense that it is well identified. This is not therefore an ambient noise of unidentified origin.

[0047] The recording of these seismic waves is carried out by means of at least one pair of sensors disposed in such a manner as to be aligned with the noise source during the aforementioned period of time.

[0048] The pairs of sensors may be disposed according to various arrangements, such as for example regularly spaced along straight lines or rows as shown in FIG. 2 described hereinafter, but they may also be spaced in an irregular fashion, following all kinds of curves, or on regular or irregular grids of square or hexagonal or any given shape. They may also be placed randomly.

[0049] FIG. 2 gives, by way of non-limiting example, a recording configuration in the case where the target zone is a railroad track 20. Nevertheless, the invention is just as applicable to a target zone consisting of a road, or else a building or another item of infrastructure, such as for example an airport.

[0050] A plurality of receivers or seismic wave sensors 22 is deployed along the railroad track 20, referred to as target track, on either side of the latter. These sensors 22 are distributed between a “Receiver row A” 22A and a “Receiver row B” 22B. The sensors 22 may for example comprise at least one geophone and/or 10 at least one accelerometer and/or at least one sensor based on the use of optical fibers, for example of the DAS (Distributed Acoustic Sensing) type.

[0051] On either side of the target railroad track 20, the sensors 22 may for example be disposed at regular intervals from one another, for example at a few meters' interval, leaving for example 2 meters between each sensor 22.

[0052] The sensors 22 may be placed on the surface of the ground or be slightly buried, in other words situated a few meters under the surface of the ground. They may be deployed on or under the ballast, or else on or under the rails or else in a railroad tunnel.

[0053] The plurality of sensors 22 comprises at least one pair of sensors 220, 221 disposed opposite each other, on either side of the railroad track 20, in a transverse manner to the latter.

[0054] The mobile seismic noise source is a train 24 travelling on a railroad track 21, referred to as source track, distinct from the target railroad track 20. The noise emitted by the moving train 24 is symbolized in the drawing by arrows in the form of circles 26. This noise mainly comes from the contact of the wheels of the train 24 on the rails of the railroad track 21.

[0055] The assembly comprising the plurality of sensors 22, the mobile source 24 of seismic noise and a module designed to carry out the processing operations of the method for monitoring the subsoil described here, forms a device for monitoring the subsoil according to the invention.

[0056] Considering the pair of sensors 220, 221, this will be located in the alignment of the train 24 for a predetermined period of time defined by a first moment in time where the point of contact of the first wheel of the train 24 with the rail is on the straight line passing through the barycenter of the sensors 220 and 221 and a second moment in time where the point of contact of the last wheel of the train 24 with the rail is on the straight line passing through the barycenter of the sensors 220 and 221.

[0057] According to the invention, during this period of time, the sensors 220 and 221 record seismic waves which will undergo later processing operations of the method according to the invention. This period of time is predetermined, given that the kinematic parameters of the train 24 are known, namely, its position and its speed at all times. In this sense, the mobile seismic noise source constituted by the train 24 is identified.

[0058] The recording step 10 may be limited to this predetermined period of time. As a variant, the recording may take place continuously over time.

[0059] With reference to the flow diagram in FIG. 1, optionally, the step 10 for recording seismic waves may be followed by a step 11 for pre-processing the seismic waves recorded by the sensors 22.

[0060] These pre-processing operations may be of various types and include at least one denoising phase and one frequency filtering phase.

[0061] In the non-limiting example of a railroad track described here, during the phase for denoising the recorded seismic waves, the main noise surrounding the seismic noise produced by the train 24 is removed. This main surrounding noise is an electrical noise associated with the overhead catenary cables whose characteristics are known. A signal with an improved signal-to-noise ratio is thus obtained.

[0062] During the frequency filtering phase, the signal is filtered within a predetermined band of frequencies which corresponds to both the estimated emission band of the train 24 and to the frequency band of interest, which depends on the depth of interest of the investigation of the subsoil under the target zone constituted by the railroad track 20.

[0063] Furthermore, between the denoising phase and the frequency filtering phase, optionally, a spectral whitening operation and/or a binarization operation may furthermore be performed.

[0064] The spectral whitening operation, also referred to as spectral equalization, a technique known per se, consists in weighting all the frequency components of the signal in such a manner that they all have the same representation in energy.

[0065] The binarization operation, also known per se, consists in assigning a value of −1 or +1 to each instantaneous value of the signal in order to simplify the signal.

[0066] Other types of pre-processing operations may be applied.

[0067] Following the recording step 10 and, potentially, the step 11 for pre-processing of the seismic waves recorded by the sensors 22, a step 12 for reconstructing a seismogram of the target zone is carried out.

[0068] This reconstruction step 12 uses the seismic waves, recorded and potentially pre-processed, and applies an interferometry technique to them.

[0069] This interferometry technique comprises, for example, at least one operation for intercorrelation of signals coming from the seismic waves respectively recorded by the sensors 220 and 221 during the predetermined period of time when the train 24 passes over the source track 21. This interferometry technique allows virtual transverse propagation paths to be reconstructed, here perpendicular to the target railroad track 20. These virtual propagation paths characterize the seismogram.

[0070] FIG. 3 illustrates schematically these virtual propagation paths for each pair of sensors opposite each other on either side of the target railroad track 20.

[0071] As a variant, any other interferometry technique could be applied during this reconstruction step 12, such as for example a method based on convolution operations instead and in place of the aforementioned intercorrelation operations.

[0072] As illustrated by FIG. 4, the reconstruction step 12 may be carried out for various orientation directions. Indeed, intercorrelation operations may be carried out, not only between the signals coming from the seismic waves recorded by the pairs of sensors situated opposite each other like the pair of sensors 220 and 221, but also between the pairs of sensors disposed in a similar manner to the sensors A5 and B6 or to the sensors A7 and B6, in other words the pairs of sensors for which the straight line passing through the barycenters of the two sensors of the pair forms a non-zero angle with the perpendicular to the target railroad track 20.

[0073] Furthermore, as illustrated in FIG. 5, the reconstruction step 12 may be carried out, not only for various orientation directions but furthermore by using several railroad tracks associated with pairs of sensors. The source track 21, parallel to the tracks 201 to 204, is situated opposite the “receiver row 22A”, as in FIGS. 2 to 4 but, for simplification, is not shown in FIG. 5.

[0074] Thus, when a train passes over the source track 21, virtual propagation paths between the rows 22A and 22B and/or between the rows 22B and 22C and/or between the rows 22C and 22D and/or between the rows 22D and 22E may be reconstructed by interferometry. In other words, for each pair of adjacent railroad tracks, each railroad track allows a reconstruction of a seismogram of the zone situated under the adjacent railroad track, using the pairs of sensors deployed on either side of this adjacent railroad track.

[0075] The reconstruction of transverse virtual propagation paths, in other words transverse to the railroad track, has been described hereinabove using the time window in which the train 24 is aligned with a pair of sensors of the type of the pair 220, 221: the pairs of sensors allow the variation of the seismic noise path to be studied.

[0076] As a variant, it is possible to use the correlation, by convolution, between two sensors on either side of a track and the signal emitted by a train on this same track. Thus, the time variations of the propagation times may be used as a continuous monitoring for the purpose of measuring any appearance of anomalies.

[0077] With reference to FIG. 1, following the reconstruction step 12, optionally, a step 13 may be carried out for addition, or “stacking”, of reconstructed signals coming from sensed seismic waves originating from several mobile sources 24 of seismic noise.

[0078] In the non-limiting example described here where these noise sources are trains, several trains travelling at different times, having different lengths and speeds, which will therefore be aligned with a pair of sensors of the type of the pair 220, 221 during different periods of time, will supply similar reconstructed signals.

[0079] The addition of these multiple signals allows the signal-to-noise ratio to be improved and hence, in the end, an image of better quality of the subsoil under the target zone to be obtained. Furthermore, for a given pair of sensors of the type of the pair 220, 221, this allows virtual propagation paths to be produced as a one-off or periodically with a chosen periodicity, namely, daily or weekly or else monthly, for example.

[0080] As shown in FIG. 1, following the reconstruction step 12 and the potential stacking step 13, a step 14 for generating an image of the subsoil under the target zone is carried out, using the previously reconstructed seismogram.

[0081] During this step 14 for generating an image, the processing operations will differ depending on the type of seismic waves recorded at the step 10. This is because either surface waves, or refracted waves, or potentially reflected waves, may be recorded.

[0082] In one embodiment where, at the step 10, seismic surface waves are recorded, the following operations are carried out in order to obtain an image of the subsoil under the target zone.

[0083] First of all, dispersion curves of the surface waves are determined in the reconstructed signals, within a predetermined band of frequencies. By way of non-limiting example, this band of frequencies may be the band from 1 to 100 Hz.

[0084] Optionally, only the maxima of these dispersion curves may subsequently be selected.

[0085] Then, using these dispersion curves, a tomography of the speed of the reconstructed signals is carried out.

[0086] Lastly, an inversion of the surface wave speeds of the reconstructed signals coming from the surface waves is carried out, so as to produce a speed model of the pressure and shear waves, known to those skilled in the art by the terminology P- and S-waves.

[0087] By virtue of the repetition of these operations over time, an image of the subsoil under the target zone is thus obtained which evolves over time, in other words, a representation in four dimensions of this region of the subsoil.

[0088] In another embodiment, where, at the step 10, refracted seismic waves are recorded, the following operations are carried out in order to obtain an image of the subsoil under the target zone.

[0089] First of all, the arrival times are determined of the reconstructed signals of the seismogram, corresponding to the refracted seismic waves.

[0090] Then, a tomography of the reconstructed signals corresponding to the refracted P-waves is carried out based on these arrival times, so as to produce a speed model of the pressure waves, known as P-waves.

[0091] In a similar manner to the embodiment where seismic surface waves are recorded, by virtue of the repetition of these operations over time, an image of the subsoil under the target zone is thus obtained which evolves over time, in other words, a representation in four dimensions of this region of the subsoil.

[0092] In yet another particular embodiment, reflected seismic waves could be recorded at the step 10. This would then lead to an image of the reflectivity of the subsoil under the target zone being obtained.

[0093] It is possible to naturally cumulate the recordings of the refracted waves with the surface waves, which accordingly increases the resolution, just as does the integration of the processing of the reflected waves into the preceding methods. In particular, the signals are processed independently up to the centers, which are subsequently combined, the centers being the dispersion curves for the surface waves and the time of first arrival for the refracted waves.

[0094] In particular, preferably and such as illustrated in FIG. 6, the image of the target subsoil 3 under a railroad track 20, of width of the order of 1.5 m, spaced from a parallel source track 21, may be formed by two rows of sensors of the conventional accelerometer type spaced by 5 m flanking the railroad track 20, the sensors 22 of each row being separated by 3 m, the sensors being disposed in a staggered manner in order to homogenize the resolution of the three-dimensional image obtained by intercorrelation so as to cover the whole track. The processing of the refracted surface waves allows a resolution of 2 m, or even less, to be attained. If a higher resolution is required, the rows of sensors may be brought closer together, for example by directly associating the sensors with the rails of the track, in other words by spacing them out by the order of 1.5 m, and/or by using two optical fibers for DAS, thus creating sensors spaced out by the order of 60 cm.

[0095] It is of course possible to add to the method according to the invention a reconstruction similar to those described in the prior art, in which groups of sensors allow the dispersion curves of the seismic waves to be analyzed, and the resolution and/or the reliability of the image to be further enhanced: virtual longitudinal propagation paths, in other words parallel to the railroad track, are thus reconstructed for example by using time windows of predetermined duration preceding and following the passage of the train 24, respectively after the sensors begin to receive seismic waves and before the sensors stop receiving seismic waves.

[0096] The method according to the invention comprises a passive monitoring of the subsoil under a target zone in the sense that it processes seismic waves received, but does not produce any. Nevertheless, it may be envisioned to combine the sensors that receive the seismic waves with active seismic sources, in order to apply processing operations of the type described in the document US-A-2017 052269.