Geophysical method for detecting soil heterogeneity by analyzing the behavior of backscattered S-waves.

Abstract

Geophysical method for detecting heterogeneity in soil, comprising the steps of distributing geophones (1) between a first end (2.1) and a second end (2.2) of at least one measurement line (2), recording the position of the geophones (1), emitting into the soil at least one incident seismic wave from a seismic source (4) in the vicinity of the first end (2.1) of the measurement line (2), recovering a measurement signal transmitted by the geophones (1).The process includes the steps of isolating representative measurement signals of backscattered shear seismic waves received in return from the incident seismic wave, processing the measurement signals to numerically quantify the intensity of the backscattered shear seismic waves in order to detect soil heterogeneity, and measuring the arrival time of a peak in intensity of the backscattered shear seismic waves in order to determine, using a shear wave velocity distribution model in the soil, the position of the detected heterogeneity. ABRIDGED FIGURE: [Fig. 1].

Description

Title of the invention: Geophysical method for detecting soil heterogeneity by analyzing the behavior of backscattered S waves.

[0001] The present invention relates to the field of geophysics and more particularly to the detection of soil heterogeneities. The invention relates to heterogeneities ranging in size from meters to several meters and of varying depths, from a few meters to several tens of meters.

[0002] BACKGROUND OF THE INVENTION

[0003] Natural and man-made cavities present significant risks in the fields of geotechnics and natural hazards. Natural cavities, such as caves, karst formations, and sinkholes, can cause collapses, landslides, and subsidence. Man-made cavities, such as old mines, quarries, tunnels, pits, or caves, can also cause collapses and ground movements, as well as foundation stability problems.

[0004] These risks are often amplified by urbanization and the expansion of infrastructure (roads, dikes, railway lines for example), which can exacerbate the effects of ground movements.

[0005] To minimize them, it is important to map and monitor soil heterogeneities such as natural and anthropogenic cavities, decompaction, concrete blocks, underground networks... and to take preventive measures, such as soil consolidation, wall stabilization and groundwater management.

[0006] Geophysical methods for detecting cavities are known, consisting of emitting seismic waves into the ground from a surface source and then measuring, from the surface, the waves backscattered towards it. Indeed, cavities have the property of interacting strongly with seismic waves by diffracting them and thus causing backscattering of a portion of them towards the source. This backscattering takes the form of a hyperbola, the visual study of which on a seismogram will reveal the presence of cavities. This approach nevertheless has several disadvantages because several phenomena combine near the surface and can alter or even mask this hyperbola, for example: - the presence of high-intensity ambient seismic noise (anthropogenic or natural) masking the backscatter hyperbola; - the presence of lower density terrain attenuating the waves as they pass through and significantly reducing their appearance on seismograms.

[0007] The spatial and temporal variations of these parameters also affect the correct identification and localization of the backscattering hyperbola.

[0008] It is known to perform fk filtering of the seismic signal to retain only negative velocities and then to carry out a visual analysis of the filtered signal to determine the presence or absence of diffraction and thus of a potential cavity. However, these filters can cause more or less significant distortion of the filtered signal, distortion which can limit the quality and accuracy of the visual analysis.

[0009] It follows that, although these methods are theoretically suitable for the detection of multi-meter cavities, the measurement conditions can make the return waves difficult to analyze, altering the ability to detect cavities.

[0010] These processes also remain highly dependent on the operator, which can degrade the repeatability of the results (particularly in contexts of studying the evolution of phenomena).

[0011] SUBJECT OF THE INVENTION

[0012] The invention is notably aimed at improving the detection of soil heterogeneity. Summary of the invention

[0013] For this purpose, the invention provides a geophysical method for detecting heterogeneity in soil, comprising the steps of distributing geophones between a first end and a second end of at least one measurement line, recording the position of the geophones, emitting into the soil at least one incident seismic wave from a seismic source in the vicinity of the first end of the measurement line, and recovering a measurement signal transmitted by the geophones.The process includes the steps of isolating representative measurement signals of backscattered shear seismic waves received in return from the incident seismic wave, processing the measurement signals to numerically quantify an intensity of backscattered shear seismic waves in order to detect soil heterogeneity, and measuring the arrival time of a peak intensity of backscattered shear seismic waves in order to determine, using a shear wave velocity distribution model in the soil, a position of the detected heterogeneity.

[0014] Thus, the method of the invention aims to locate heterogeneities present in the soil using the backscattering of shear seismic waves and to deduce the depth of the heterogeneities using the round-trip time of the seismic waves with respect to the heterogeneity.

[0015] Shear seismic waves are chosen over compression waves because they represent a larger share of the energy released by a seismic source (on average 30% for compressional seismic waves and 70% for shear seismic waves).

[0016] It is understood that, during the emission of an incident seismic wave, the latter is transmitted through the underlying ground. If the incident seismic wave encounters a contrast in mechanical behavior (more compact or, conversely, less compact ground), a portion of the incident seismic wave will be reflected at this interface and will "return" to the seismic source (this reflected portion of the incident seismic wave is called a "backscattered seismic wave"). The greater the contrast in mechanical behavior, the greater the amount of backscattered seismic wave. In natural environments, strong point contrasts in seismic impedance are rare. These contrasts are often linked to cavities, dissolution zones, or decompaction zones. In urban areas, these contrasts can also be linked to the passage of meter-sized utility networks, concrete blocks in embankments, etc.Therefore, in urban environments, it is more difficult to differentiate a cavity from another highly contrasting man-made feature.

[0017] The method of the invention further utilizes the arrival time of the peak intensity of the backscattered seismic wave, which corresponds to the round-trip time of the seismic waves between the source and the heterogeneity. The round-trip time of the seismic waves depends on the position (vertical and in depth) of the heterogeneity relative to the source, as well as on the distribution of the propagation velocity of the backscattered seismic waves in the encountered soils.

[0018] According to optional features, used individually or in whole or in combination: - the measurement signal from each geophone is analyzed according to time windows of increasing duration; - two successive analysis time windows have a duration difference of 5 ms; - in each analysis time window, the measurement signal is decomposed into the frequency / wavenumber spectrum and, preferably, in the frequency / wavenumber spectrum, the measurement signal is filtered according to velocities to retain only the part of the measurement signal representative of backscattered seismic shear waves; - a principal component analysis with data standardization is performed on said part of the measurement signal to obtain a maximum eigenvalue which is recorded in analysis data; - heterogeneity is considered to be present in the event of a significant decrease in the principal eigenvalue for at least two consecutive or adjacent measurement points and one of the time windows of analysis; - the process includes the step of filtering the analysis data using a spatial Fourier transform filter to attenuate spatial noise; - the process includes the step of filtering the analysis data using a regional trend filter in order to eliminate variations in the data related to a regional trend in order to retain only local variations; - regional trend filtering is performed by subtracting a polynomial regression curve from the set of analysis data to retain only the analysis data below the polynomial regression curve; - for each measurement point, the measurement signals from the geophones are used to define a spatiotemporal acquisition profile based on a distance from each geophone to the seismic source and an instant of emission of the measurement signal from each geophone to determine an arrival time of the variations in eigenvalues; - the process includes the step of using the spatiotemporal profile and a model of the distribution of propagation velocities of seismic shear waves in the ground to calculate a round-trip path length of seismic shear waves from the seismic source to each geophone, the round-trip path having a midpoint corresponding to an edge of the heterogeneity closest to the seismic source.

[0019] The method of the invention is particularly well suited to the detection of metric heterogeneities in X, Y and Z over a depth ranging from 0m to 30m and allows: - good on-site measurement efficiency; - rapid data processing; - a discrimination allowing for an effective interpretation; - on-site adjustability; - total or partial independence from the operator during processing.

[0020] Other features and advantages of the invention will become apparent from the following description of a particular, non-limiting embodiment of the invention. Brief description of the drawings

[0021] Reference will be made to the attached drawings, among which:

[0022] [Fig. 1] is a cross-sectional diagram illustrating the implementation of the method of the invention.

[0023] [Fig.2] is a flowchart of the process according to the invention. DETAILED DESCRIPTION OF THE INVENTION

[0024] With reference to [Fig. 1], the method of the invention is implemented by means of a measurement system comprising a towable device including geophones 1 distributed along a support strip 2 forming a power cable for the geophones 1 and defining a measurement line having a first end 2.1 and a second end 2.2. There are twenty-four geophones 1 spaced two by two at 0.5 m intervals, giving a strip 2 with a length of 11.5 m. The towable device is known per se and is often referred to as a streamer.

[0025] The measurement system also includes an acquisition console 3 which is arranged to be connected to the geophones 1 and includes at least one processor and a memory containing a computer program executable by the processor to implement the method of the invention which will be detailed later.

[0026] The measurement system also includes a seismic source, symbolized as 4 in [Fig. 1], to produce an incident seismic wave. The seismic source can be a mass weighing between 5 and 8 kg, which is manipulated by an operator to strike a plate resting on the ground. If it is necessary to produce a more powerful incident seismic wave (for example, due to high coactivity or excessive ambient noise), it is possible to use an accelerated falling seismic source such as the PEG-40, marketed by GEOREVA.

[0027] The method of the invention will now be described in relation to soil inspection over a predefined area.

[0028] A series of measurements will be carried out by positioning the towable device at several locations (called measurement points) within the predefined area. The measurement points are usually spaced one to two meters apart. As is often the case with spot geophysical measurements, the spacing of the measurements defines the size and depth of the heterogeneities that can be detected: the greater the spacing between measurements, the smaller and shallower the heterogeneity must be, or conversely, the larger and deeper the heterogeneity must be. A significant anomaly should ideally be defined by at least two consecutive or adjacent measurement points in order to eliminate the suspicion of a single measurement error.

[0029] By way of example, during tests, a spacing of two meters between measurement points made it possible to observe a cavity two meters wide at a depth of eight meters. In order to move the towable device from one measurement point to another, the towable device can be towed by an operator (low efficiency but with greater adaptability to the site and a smaller lateral footprint) or by a vehicle (higher efficiency but low adaptability to the site and a larger lateral footprint - i.e. the length of the vehicle).

[0030] The geographic coordinates (local reference frame or satellite positioning coordinates) of each measurement point must be entered in order to provide an accurate map of the distribution of results. To this end, the complete geometry of the system must be known: the spacing of the geophones 1 from each other and the position of the seismic source 4 relative to the geophones 1.

[0031] The location of the measurement point of the system is located at the geophone furthest from the source.

[0032] The coordinates of each measurement point must be recorded either in an absolute coordinate system (geographic coordinates) or in a local coordinate system. In both cases, the resolution of the coordinates must be consistent with the size of the heterogeneity being sought. In the present example, the method is implemented to search for heterogeneities of metric size: it is advisable that the resolution of the measured geographic coordinates be less than 50 cm

[0033] A single-point measurement will now be described.

[0034] The towable device is positioned in the predefined area by laying the strip 2 and the geophones 1 on the ground at the intended measurement point. The seismic source 4 is placed approximately one meter from the first end 2.1, while the acquisition console is placed on the side of the second end 2.2.

[0035] The strike is made by the mass on the plate resting on the ground 1 meter from the first end 2.1 of the band 2.

[0036] The acquisition console 3 is arranged to process the following data: - the seismic data provided by each geophone 1 in the form of an electrical measurement signal; - a suitable sampling frequency (generally less than 0.0625ms); - a listening time long enough to be able to capture backscattered seismic waves and short enough not to be too disturbed by ambient noise beyond the time of capture of backscattered seismic waves (generally less than 500ms).

[0037] More specifically, the computer program executed by the acquisition console 3 is arranged to record the measurement signals provided by the geophones 1 and to perform the processing described below with reference to [Fig.2].

[0038] The measurement signal produced by the set of geophones 1 is segmented into increasing time segments. For example, the entire measurement signal is segmented according to different analysis time windows ranging from 0ms to 5ms; 0ms to 10ms; 0ms to 15ms; 0ms to 20ms; ...; 0ms to 500ms. These different analysis time windows are intended to allow the detection of the arrival time of backscattered seismic shear waves. The step size between the time windows The analysis step can be modified, but the five millisecond step seems to be a good compromise between a step that is too small, which would produce aliasing, and a step that is too large, which would not allow sufficient precision on the appearance of the backscattered shear seismic wave.

[0039] For each analysis time window, the computer program executed by the acquisition console 3 is arranged to: - decompose the measurement signal into the "frequency / wavenumber" spectrum; - in the said "frequency / wavenumber" spectrum, filter the measurement signal to retain only the part representative of the velocities of seismic waves returning to the seismic source (the velocities are positive or negative depending on the geometry of the acquisition system); - perform a principal component analysis with standardization of the filtered signal data; - record only the maximum eigenvalue from this analysis.

[0040] This first processing makes it possible to obtain, for each measurement point and for all the defined time windows of analysis, the maximum eigenvalue of the principal component analysis with standardization of the "frequency / wavenumber" spectrum data of backscattered seismic ground.

[0041] The response of a soil heterogeneity is revealed, in the present process, as a strong decrease in the maximum eigenvalue at the right and in the vicinity of the heterogeneity.

[0042] This local decrease must be highlighted in the general behavior defined below.

[0043] If a linear measurement profile is defined, the set of measurements may exhibit a general behavior (or trend) related to the ground itself (i.e., a gradual decrease or increase in values ​​is linked to the terrain, without necessarily indicating the presence of anomalies). The same applies in plan view, with an overall behavior of the measurements along a preferred axis.

[0044] It is therefore necessary to differentiate between the general behavior (trend of measurements) and the abnormal variation of measurements according to this trend (local trend). This is the same principle as in microgravimetry where one wants to distinguish abnormal point behaviors within a general (standard) evolution of measurements.

[0045] The overall behaviors are derived from polynomial regressions (linear or planar depending on the case studied) ranging (from experience) from order 1 to 3 maximum and being at the estimate of the operator in charge of the data processing.

[0046] In order to highlight this local decrease in the overall behavior, the computer program of the acquisition console 3 filters the previously processed data using two main filters: - a spatial Fourier transform filter (spatial noise filtering); - a filter for regional trends.

[0047] The spatial Fourier transform filter allows the data from the distance / intensity coordinate system (from the initial processing, which provides eigenvalue data, namely intensity, as a function of the spatial distribution of the measurement points, namely distance) to the wavenumber / intensity coordinate system. It is therefore possible to filter the wavenumber from the influence of the anomalies being sought on the overall measurement response. Thus, this filter removes spurious oscillations from the data by determining the range of wavenumbers to be eliminated from the variation in the data for each profile. The spatial transform filter thus makes it possible to discriminate between irrelevant variations in the data in order to focus on a specific anomaly size.

[0048] The regional trend filter removes data variations related to a regional trend, retaining only local variations (which may be linked to heterogeneities). The regional trend filter highlights anomalies by increasing their importance relative to other measurements. For each profile, the filtering is performed by subtracting a polynomial regression curve of the entire data set from the data curve. Since the response of a heterogeneity in the soil is a sharp decrease in the maximum eigenvalue, only values ​​below the polynomial regression curve are retained.

[0049] With regard to the depth localization of the heterogeneity, the arrival time of the backscattered seismic wave will depend on the distance between the measurement system and the heterogeneity. Thus, from a line of measurement points (“acquisition profile”), it is possible to observe the variation in the arrival time of the backscattered seismic wave (arrival time of the variations in eigenvalues) from the position of the measurement system relative to the heterogeneity.

[0050] The variation in the arrival time of the backscattered seismic wave is related to the model of shear wave propagation velocities in the ground. If this model is provided, it is possible to simulate the arrival time of the backscattered shear wave for a given heterogeneity position.

[0051] The principle of the depth determination of the detected heterogeneities is as follows: for each point available in the "distance / time" frame of the profile studied, the computer program brings back to its center the energy observed if the point in question is a diffracting point.

[0052] For calculating the round-trip time of the backscattered seismic wave, experimentation shows that: - The location of seismic source 4 must be used as the origin of the forward emission. - the edge of the heterogeneity closest to the seismic source 4 must be used as the midpoint of the round trip path; - the end of the measurement profile furthest from the seismic source 4 must be used as the arrival point of the round trip path of the backscattered seismic wave.

[0053] This operation will therefore make it possible to focus, at a point in the "distance / time" coordinate system, the energy associated with the seismic wave backscattered by a heterogeneity. Given the theoretical velocity of the ground shear seismic waves, this operation thus results in focusing, in the "distance / depth" coordinate system, the location of the peak values ​​associated with the presence of heterogeneities.

[0054] It is therefore possible to develop a distribution map of the maximum eigenvalue from the principal component analysis with standardization of the frequency / wavenumber spectrum data of the backscattered seismic wave. Areas of strong negative variations indicate the presence of heterogeneities in the soil (cavities, significant decompaction, anthropogenic features such as concrete blocks or networks of metric dimensions). The processing thus results in a two-dimensional horizontal (xy) output at this level.

[0055] These anomalies can then be analyzed to determine their depth from an initial model of the propagation velocity distribution of shear seismic waves in the ground. The processing thus results in a three-dimensional (xyz) result.

[0056] Of course, the invention is not limited to the embodiment described but encompasses any variant falling within the scope of the invention as defined by the claims.

[0057] In particular, the measurement system may have a different structure from that described.

[0058] The incident seismic wave is produced by the impact of a mass on a rigid plate placed on the ground. This mass can be manipulated directly by an operator or mounted in a frame to support the mass above the rigid plate and to release the mass from a predefined height at the operator's command.

[0059] The analysis time windows may have a different duration than that mentioned.

[0060] Geophones can be separated from each other by a shorter or longer distance.

[0061] It should be noted that, during the FK transform, for a seismogram including all the relevant traces (with or without equivalent frequency filtering on all the traces), it is the shear waves that stand out in the diagram because shear waves represent 70% of the energy. Compression waves are negligible.

[0062] It is also possible to do without all the filters in order to study all the data.

Claims

Demands

1. Geophysical method for detecting heterogeneity in soil, comprising the steps of distributing geophones (1) between a first end (2.1) and a second end (2.2) of at least one measurement line (2), recording the position of the geophones (1), emitting into the soil at least one incident seismic wave from a seismic source (4) in the vicinity of the first end (2.1) from the measurement line (2), recover a measurement signal transmitted by the geophones (1), characterized in that the method comprises the steps of isolating the measurement signals representative of the backscattered seismic shear waves received in return from the incident seismic wave, processing the measurement signals to digitally quantify an intensity of the backscattered seismic shear waves in order to detect a heterogeneity of the soil, and measuring the arrival time of a peak in intensity of the backscattered seismic shear waves in order to determine, using a model of the velocity distribution of the seismic shear waves in the soil, a position of the detected heterogeneity.

2. A method according to claim 1, wherein the measurement signal of each geophone (1) is analyzed according to analysis time windows of increasing durations.

3. A method according to claim 1 or 2, wherein two successive analysis time windows have a duration difference of 5 ms.

4. A method according to claim 1 or 2, wherein, in each analysis time window, the measurement signal is decomposed into the frequency / wavenumber spectrum.

5. A method according to claim 4, wherein, in the frequency / wavenumber spectrum, the measurement signal is filtered as a function of velocities to retain only the part of the measurement signal representative of backscattered shear seismic waves.

6. A method according to claim 5, wherein a principal component analysis with data standardization is performed on said part of the measurement signal to obtain a maximum eigenvalue which is recorded in analysis data.

7. A method according to claim 6, wherein heterogeneity is considered to be present in the event of a significant decrease in the principal eigenvalue for at least two consecutive or adjacent measurement points and one of the analysis time windows.

8. A method according to claim 7, comprising the step of filtering the analysis data using a spatial Fourier transform filter to attenuate spatial noise.

9. A method according to claim 7 or 8, comprising the step of filtering the analysis data using a regional trend filter in order to eliminate variations in the data related to a regional trend in order to retain only local variations.

10. A method according to claim 9, wherein regional trend filtering is performed by subtracting a polynomial regression curve from the set of analysis data to retain only the analysis data below the polynomial regression curve.

11. A method according to any one of claims 6 to 10, wherein, for each measurement point, the measurement signals from the geophones (1) are used to define a spatiotemporal acquisition profile as a function of a distance from each geophone (1) to the seismic source (4) and an emission time of the measurement signal from each geophone (1) to determine an arrival time of the eigenvalue variations.

12. A method according to claim 11, comprising the step of using the spatiotemporal profile and a seismic shear wave propagation velocity distribution model in the ground to calculate a round-trip path length of the seismic shear waves from the seismic source (4) to each geophone (1), the round-trip path having a midpoint corresponding to an edge of the heterogeneity closest to the seismic source (4).

Images