Joint imaging method of multi-dimensional seismic background noise in underground space

Abstract

The present invention discloses a method for joint imaging of multi-dimensional seismic background noise in underground space, comprising the following steps: collecting seismic background noise data and performing data preprocessing; performing layer stripping imaging to obtain a geological structure model and geological stratification data; combining the geological stratification results obtained from the layer stripping imaging with surface wave dispersion curve inversion to obtain an underground shear wave velocity field; using the dispersion curve inversion results as the initial input model and the geological structure obtained from the layer stripping imaging as the constraint term, constructing a full waveform inversion target universal function, calculating the shear wave velocity gradient, performing iterative updates, and outputting the final shear wave velocity field when the new velocity model meets the convergence conditions. Through multi-dimensional background noise imaging technology, in response to the increasing complexity of the geological surface environment, full coverage surveys of different spatial continuity areas such as mountainous areas, cities, and plains can be achieved, with more accurate results.

Description

Technical Field

[0001] The present invention relates to the field of engineering geophysical non-destructive detection technology, and in particular to a method for joint imaging of multi-dimensional seismic background noise in underground space. Background Art

[0002] In geological surveys for projects like high-speed rail and highway construction and urban underground space development and utilization, the precise characterization of shallow surface geological structures and anomalous targets (karst, air-raid shelters, unevenly weathered bodies, groundwater seepage channels, weak interlayers, etc.) is crucial for construction safety and presents a critical technical challenge for construction managers. Seismic background noise imaging technology records what is traditionally considered "seismic noise" and converts it into "effective" information. It can adaptively address interference from human activity, electromagnetic fields, and vehicles. It eliminates the need for artificial seismic sources and offers advantages such as repeatability, environmental friendliness, and high timeliness. It is currently one of the most popular techniques for capturing the structure of underground media.

[0003] However, traditional single seismic background noise imaging technology has varying degrees of defects in terms of resolution and applicability: 1) The 1D method has good spatial applicability, but can only provide geological structure information and cannot obtain medium properties; 2) Although 1.5D dispersion curve inversion can obtain shear wave velocity properties, due to the equivalence of lateral acquisition, the lateral resolution of the output results is difficult to guarantee; 3) 2D and 3D background noise full waveform inversion imaging (Full Wave Inversion) technology based on wave equation theory can provide an inversion result that takes into account both vertical and lateral resolution, but the FWI method is a model parameter reverse reconstruction technology with extremely strong nonlinear properties. It is very easy to fall into local extreme values ​​when the initial input model is inaccurate and the data frequency band is missing.

[0004] To this end, the present invention combines the advantages of background noise imaging methods in various dimensions and proposes a joint imaging method for multi-dimensional seismic background noise in underground space. This method fully mines the geological attribute information contained in the seismic background noise signal from different dimensions. It can not only achieve full coverage survey of different spatial continuity areas, but also complete the mutual integration and verification of multi-dimensional seismic noise survey results, thereby significantly improving the accuracy of underground space detection. Summary of the Invention

[0005] Therefore, the purpose of the present invention is to provide a method for joint imaging of multi-dimensional seismic background noise in underground space, which adopts 1D layer peeling imaging, 1.5D dispersion curve inversion and 2D waveform inversion. In response to the increasing complexity of the geological surface environment, multi-dimensional seismic background noise joint imaging technology is adopted to achieve full coverage survey of different spatial continuity areas.

[0006] To achieve the above objectives, the present invention provides a method for joint imaging of multi-dimensional seismic background noise in underground space, comprising the following steps:

[0007] S1. Node-type seismographs are evenly spaced in a pre-set linear observation system to collect original seismic background noise data.

[0008] S2. Preprocessing the original seismic background noise data, performing layer stripping imaging based on the preprocessed data, and obtaining a geological structure model and geological layering data;

[0009] S3. Based on the preprocessed data and the defined virtual shot-detector position relationship, the empirical Green's function is extracted using interferometry technology to calculate the dispersion spectrum. The surface wave dispersion curve is picked according to the dispersion spectrum trend. The surface wave dispersion curve is combined with the geological layering data obtained by layer stripping imaging to perform dispersion curve inversion to obtain the underground shear wave velocity field;

[0010] S4. Data match the virtual shot gather records with the synthetic shot gather records, use the geological structure model obtained in S2 as a constraint condition, construct the full waveform inversion target function, calculate the gradient of the shear wave velocity, update the current velocity model according to the iterative step size, and obtain a new velocity model; determine whether the iteratively updated velocity model meets the convergence condition; if the convergence condition is met, output the final shear wave velocity field; if the convergence condition is not met, repeat step S4 using the updated shear wave velocity model.

[0011] Further preferably, in S2, the method of preprocessing the original seismic background noise data includes: band-pass filtering and One-Bit normalization.

[0012] Further preferably, in S2, performing layer stripping imaging based on the preprocessed data to obtain the geological structure model and geological layering data comprises the following steps:

[0013] S201, calculating the frequency spectrum and spectral ratio curve of the seismic background noise data at each acquisition point;

[0014] S202: extracting peak values ​​from the spectral ratio curve to obtain dominant frequencies; and arranging the frequencies in descending order.

[0015] S203. Perform layer stripping imaging based on the relationship between the dominant frequency and the stratum thickness to obtain a geological structure model, stratum stratification, and thickness information of each layer.

[0016] Further preferably, in S203, the relationship between the dominant frequency and the formation thickness is calculated according to the following formula:

[0017]

[0018] in, is the thickness of the nth layer, is the velocity of the nth layer, is the predominant frequency of the nth stratum.

[0019] Further preferably, in S3, the inversion of the dispersion curve is performed using the surface wave dispersion curve in combination with the geological stratification data obtained by layer stripping imaging to obtain the underground shear wave velocity field, including using the obtained stratigraphic stratification and thickness information of each layer as constraints, and performing the dispersion curve inversion according to the following formula:

[0020]

[0021] in, is the objective function of dispersion curve inversion, is the actual observed dispersion curve, is the dispersion curve obtained by forward calculation.

[0022] Further preferably, in S4, the virtual shot gather record is obtained by convolution calculation using a given wavelet signal and an extracted empirical Green's function; the synthetic shot gather record is obtained by forward modeling the wave equation based on a given seismic wavelet using the underground shear wave velocity field obtained by inversion of the dispersion curve as the initial velocity model.

[0023] Further preferably, in S4, a full waveform inversion target universal function is constructed according to the following formula to obtain a new velocity model;

[0024]

[0025] in, Synthetic shot gather records generated for forward simulation, For virtual gun collection records, T is the length of time recorded, is the constraint weight coefficient, and m is the geological structure model.

[0026] The present invention discloses a method for joint imaging of multi-dimensional seismic background noise in underground space, which has at least the following advantages over the existing technology:

[0027] (1) Strong spatial applicability.

[0028] Multi-dimensional seismic background noise imaging technology includes three imaging methods: 1D layer peeling imaging, 1.5D dispersion curve inversion, and 2D waveform inversion. In response to the increasing complexity of the geological surface environment, multi-dimensional background noise imaging technology can achieve full coverage surveys of different spatial continuity areas, such as mountainous areas (applicable to 1D methods), cities (applicable to 1D and 1.5D), and plains (applicable to 1D, 1.5D, and 2D).

[0029] (2) The detection accuracy is reliable.

[0030] In the multi-dimensional seismic background noise joint imaging technology, the mutual fusion constraints between multi-dimensional imaging results can further improve the comprehensive interpretation accuracy of underground spatial structures. Compared with the traditional single seismic background noise imaging technology, multi-dimensional joint inversion has obvious advantages in accuracy and precision. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a structural schematic diagram of a method for joint imaging of multi-dimensional seismic background noise in underground space provided by the present invention.

[0032] Figure 2 This is the 1D layer peeling imaging result.

[0033] Figure 3 This is the 1.5D dispersion curve inversion result diagram.

[0034] Figure 4 This is the 2D waveform inversion result diagram. DETAILED DESCRIPTION

[0035] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0036] like Figure 1 As shown, the method for joint imaging of multi-dimensional seismic background noise in underground space provided by an embodiment of one aspect of the present invention includes the following steps:

[0037] S1. Target points are evenly spaced within the pre-set linear observation system. Raw seismic background noise data is collected at these target points. Seismographs placed at these target points are then blindly recorded for a specified period of time. This period typically requires at least 30 minutes of noise recording. This period can be extended to ensure a high signal-to-noise ratio.

[0038] S2. Preprocessing the original seismic background noise data, performing layer stripping imaging based on the preprocessed data, and obtaining a geological structure model and geological layering data;

[0039] The pre-processing methods include: band-pass filtering and one-bit normalization. (1) Band-pass filtering is determined based on the project detection depth and the background noise frequency band, generally 5-30Hz is the effective frequency band range; (2) One-bit normalization is mainly used to enhance the phase consistency between multiple signals:

[0040] Obtaining geological stratification data and geological structure models specifically includes the following steps:

[0041] S201, calculate the frequency spectrum and spectral ratio curve of the seismic background noise data at each acquisition point;

[0042] S202: Extract the peak value in the spectrum ratio curve to obtain the dominant frequency f; arrange the dominant frequencies in descending order ( f1 , A1 )、( f2 , A2 )、( f3 , A3 )…( fn , An );

[0043] S203. Perform layer stripping imaging based on the relationship between the dominant frequency and stratum thickness as shown below to obtain a geological structure model, stratum stratification, and thickness information of each layer.

[0044]

[0045]

[0046] Formula (1)

[0047]

[0048]

[0049] in, is the thickness of the nth layer, is the velocity of the nth layer, is the dominant frequency of the nth formation. It should be noted here that this velocity is generally the approximate velocity range of the detection area obtained through geological mapping or drilling, with low accuracy requirements, and is different from the velocity field obtained by subsequent dispersion curve inversion and waveform inversion. A Differentiate and map different strata, and then interpret them to obtain geological structure models m ;

[0050] S3. Define the positional relationship between virtual shot points and receiver points in the linear observation system. Based on the preprocessed data, use interferometry to extract the empirical Green's function, calculate the dispersion spectrum, and pick up the surface wave dispersion curve based on the dispersion spectrum trend. Use the surface wave dispersion curve combined with the geological layering data obtained by layer stripping imaging to perform curve inversion to obtain the underground shear wave velocity field.

[0051] With the obtained stratigraphic layers and the thickness of each layer as constraints, the dispersion curve inversion is performed according to the following formula:

[0052]

[0053] in, is the objective function of dispersion curve inversion, is the actual observed dispersion curve, is the dispersion curve obtained by forward calculation.

[0054] The above formula is the objective function of dispersion curve inversion, where is the actual observed dispersion curve, This is the dispersion curve obtained by forward calculation. The inversion parameters are mainly the thickness of each layer. and speed value The dispersion curve inversion feature of the present invention is that the thickness of each layer obtained from 1D layer peeling imaging is used to obtain the thickness of each layer. h , approximate formation velocity value v and the number of strata n As a constraint, the search range of dispersion inversion is further narrowed and the stability is improved;

[0055] S4. Data match the virtual shot gather records with the synthetic shot gather records, use the geological structure model obtained in S2 as a constraint condition, construct the full waveform inversion target function, calculate the gradient of the shear wave velocity, update the current velocity model according to the iterative step size, and obtain a new velocity model; determine whether the iteratively updated velocity model meets the convergence condition; if the convergence condition is met, output the final shear wave velocity field; if the convergence condition is not met, repeat step S4 using the updated shear wave velocity model.

[0056] The virtual shot gather record is obtained by convolution calculation using a given wavelet signal and the extracted empirical Green's function; the synthetic shot gather record is obtained by forward modeling the wave equation based on the seismic wavelet obtained from the observation record, using the underground shear wave velocity field obtained by inversion of the dispersion curve as the initial model.

[0057] In S4, the full waveform inversion objective function is constructed according to the following formula to obtain a new velocity model;

[0058]

[0059] in, Synthetic shot gather records generated for forward simulation, For virtual gun collection records, T is the length of time recorded, is the constraint weight coefficient, and m is the geological structure model.

[0060] S401, deriving a gradient formula of the objective function with respect to the shear wave velocity according to the adjoint state method, and calculating the gradient field;

[0061] S402, calculating the iteration step size and updating the current velocity model;

[0062] S403, determining whether the iteratively updated velocity model meets the convergence condition; if so, executing step 404; if not, repeating steps S401-S402 using the updated shear wave velocity model;

[0063] S404: After the program is executed, the final shear wave velocity field is output.

[0064] In order to test the application effect of the method of the present invention, we used the background noise data actually observed in a certain work area to verify the effectiveness and reliability of the multi-dimensional seismic background noise joint imaging method.

[0065] Background noise data was observed using 30 node-type digital seismometers (operating in a rolling survey) spaced 2 meters apart. The raw data were first preprocessed using bandpass filtering (with an effective frequency range of 5-30 Hz) and one-bit normalization to enhance the co-phase between the multi-channel signals.

[0066] The background noise data of each measuring point is extracted, and spectrum calculation, spectrum ratio calculation and 1.5D layer peeling imaging are carried out in sequence. Figure 2 is the layer peeling imaging result, where the amplitude in the image is the relative amplitude of the dominant period, which can reflect the stiffness or fragility of the formation to a certain extent; then, the stratified results of the layer peeling imaging are constrained to perform 1.5D dispersion curve inversion, and the shear wave velocity inversion results are as follows: Figure 3 As shown; Finally, the dispersion curve inversion results ( Figure 3 ) is the initial input model, and the layer peeling imaging result ( Figure 2 ) is used as the construction constraint term to perform 2D waveform inversion. The shear wave velocity inversion results are as follows: Figure 4 As shown. In general, the results of the three methods are roughly similar, and they can all reflect the shape and position of underground anomalies. Moreover, as the dimension of the methods continues to increase, the imaging accuracy also continues to increase; the mutual fusion and verification between the multi-dimensional seismic noise survey results also increase the reliability of the underground space detection results. The multi-dimensional seismic background noise imaging technology described in the present invention mainly includes three imaging methods: 1D layer peeling imaging, 1.5D dispersion curve inversion and 2D waveform inversion. In the subsequent implementation process, one of the methods or multiple methods can be selectively carried out for different geological surface environments. Joint imaging, thereby achieving full coverage surveys of different spatial continuity areas such as mountainous areas (1D method is applicable), cities (1D and 1.5D are applicable), and plains (1D, 1.5D, and 2D are all applicable).

[0067] Obviously, the above embodiments are merely examples for clarity of explanation and are not intended to limit the implementation methods. Those skilled in the art will readily appreciate that other variations or modifications based on the above descriptions are possible. It is not necessary and impossible to enumerate all implementation methods here. Obvious variations or modifications arising therefrom remain within the scope of protection of the present invention.

Claims

1. A method for joint imaging of multi-dimensional seismic background noise in underground space, characterized in that: include: S1. Node-type seismographs are evenly spaced in a pre-set linear observation system to collect raw seismic background noise data. S2. Preprocessing the original seismic background noise data, performing layer stripping imaging based on the preprocessed data, and obtaining geological layering data and a geological structure model; S3. Based on the preprocessed data and the defined virtual shot-detector position relationship, the empirical Green's function is extracted using interferometry technology to calculate the dispersion spectrum. The surface wave dispersion curve is picked according to the trend of the dispersion spectrum. The surface wave dispersion curve is combined with the geological layering data obtained by layer stripping imaging to perform dispersion curve inversion to obtain the underground shear wave velocity field. The dispersion curve inversion is performed on the geological layering data obtained by layer stripping imaging to obtain the underground shear wave velocity field, including: in, is the objective function of dispersion curve inversion, is the actual observed dispersion curve, is the dispersion curve obtained by forward calculation; S4. Data match the virtual shot gather records with the synthetic shot gather records, use the geological structure model obtained in S2 as a constraint term, construct a full waveform inversion target universal function, calculate the gradient of the shear wave velocity, update the current velocity model according to the iterative step size, and obtain a new velocity model; determine whether the iteratively updated velocity model meets the convergence conditions; if the convergence conditions are met, output the final shear wave velocity field; if the convergence conditions are not met, repeat step S4 using the updated shear wave velocity model.

2. The underground space multi-dimensional seismic background noise joint imaging method according to claim 1, characterized in that: In S2, the method of preprocessing the original seismic background noise data includes: bandpass filtering and one-bit normalization.

3. The underground space multi-dimensional seismic background noise joint imaging method according to claim 1, characterized in that: In S2, performing layer stripping imaging based on the pre-processed data to obtain geological layering data and a geological structure model includes the following steps: S201, calculating the frequency spectrum and spectral ratio curve of the seismic background noise data at each acquisition point; S202: extracting peak values ​​from the spectral ratio curve to obtain dominant frequencies; and arranging the frequencies in descending order. S203. Perform layer stripping imaging based on the relationship between the dominant frequency and the stratum thickness to obtain a geological structure model, stratum stratification, and thickness information of each layer.

4. The underground space multi-dimensional seismic background noise joint imaging method according to claim 3, characterized in that: In S203, the relationship between the dominant frequency and the formation thickness is calculated according to the following formula: in, is the thickness of the nth layer, is the velocity of the nth layer, is the predominant frequency of the nth stratum.

5. The underground space multi-dimensional seismic background noise joint imaging method according to claim 1, characterized in that: In S4, the virtual shot gather record is obtained by convolution calculation using a given wavelet signal and an extracted empirical Green's function; the synthetic shot gather record is obtained by forward modeling the wave equation based on a given seismic wavelet using the underground shear wave velocity field obtained by inversion of the dispersion curve as the initial velocity model.

6. The underground space multi-dimensional seismic background noise joint imaging method according to claim 4, characterized in that: In S4, the full waveform inversion objective function is constructed according to the following formula to obtain a new velocity model; in, Synthetic shot gather records generated for forward simulation, For virtual gun collection records, T is the length of time recorded, is the constraint weight coefficient, and m is the geological structure model.

Images