Method and system for constructing three-dimensional fine velocity structure model of complex engineering site area

Abstract

This invention relates to the field of engineering geological exploration technology, specifically to a method and system for constructing a three-dimensional fine velocity structure model for complex engineering sites. The method first constructs a one-dimensional initial velocity structure model based on borehole data from the engineering site, including P-wave velocity, S-wave velocity, and the P-wave / S-wave velocity ratio. Then, a three-dimensional absolute P-wave velocity model is obtained through active source body wave exploration and converted into an initial S-wave velocity model. Simultaneously, active source high-frequency surface wave dispersion data, passive source surface wave dispersion data, and single-station HVSR curves are acquired. Finally, using the one-dimensional initial velocity structure model as a reference and the initial S-wave model as the initial inversion model, multi-source data are integrated for joint inversion to obtain a three-dimensional fine velocity structure model. This method effectively reduces the ambiguity of inversion results, achieves high-precision modeling across the entire depth range of complex sites, and provides reliable support for seismic safety assessment of major engineering sites.

Description

Technical Field

[0001] This application relates to the field of engineering geological exploration technology, specifically to a method and system for constructing a three-dimensional fine velocity structure model of a complex engineering site. Background Technology

[0002] In earthquake prevention and disaster reduction work for major infrastructure projects such as hydropower and nuclear power projects, a three-dimensional fine velocity structure model of the project site area is the core foundation for site seismic safety evaluation. Its accuracy and reliability determine the rationality of ground motion input, and thus affect the safety and economy of the project's seismic design. Currently, among the velocity structure modeling technologies for project sites, pure borehole wave velocity testing has the problems of limited coverage and high implementation cost. Conventional active source exploration is difficult to achieve three-dimensional continuous modeling of the entire site. Pure passive source background noise imaging lacks absolute velocity constraints and has strong multiple solutions in the inversion results. Existing surface wave and HVSR joint inversion schemes have limited ability to constrain ultra-shallow strata. None of these can adapt to the high-precision modeling requirements of the entire depth range of complex project sites such as wide valleys and deep overburden layers, and cannot meet the rigid requirements of seismic fortification for major projects. Summary of the Invention

[0003] To solve, or at least partially solve, the above-mentioned technical problems, this application provides a method and system for constructing a three-dimensional fine velocity structure model of a complex engineering site.

[0004] In a first aspect, the present invention provides a method for constructing a three-dimensional fine velocity structure model of a complex engineering site, comprising the following steps: S1. Obtain borehole data from the engineering site area, and construct a one-dimensional initial velocity structure model based on the borehole data, including P-wave velocity, S-wave velocity, and P-wave / S-wave velocity ratio. S2. Active source volume wave exploration is carried out based on the dense array of stations deployed in the engineering site area. The three-dimensional absolute P-wave velocity model is obtained by inversion of the first arrival travel time tomography of the volume wave, and the three-dimensional absolute P-wave velocity model is converted into the initial S-wave velocity model. S3. Obtain the high-frequency surface wave dispersion data of the active source, the surface wave dispersion data of the passive source, and the single HVSR curve of each station in the engineering site area. S4. Using the one-dimensional initial velocity structure model as a reference and the initial S-wave velocity model as the initial inversion model, the active source high-frequency surface wave dispersion data, the passive source surface wave dispersion data, and the single HVSR curve are integrated for joint inversion to obtain a three-dimensional fine S-wave velocity structure. The final three-dimensional fine velocity structure model is then output by combining the three-dimensional absolute P-wave velocity model.

[0005] Optionally, the conversion of the three-dimensional absolute P-wave velocity model into an initial S-wave velocity model specifically includes: For areas within the engineering site covered by the borehole data, the P-wave velocity corresponding to the three-dimensional absolute P-wave velocity model is converted into S-wave velocity using the measured P-wave velocity to S-wave velocity ratio from the borehole data. For unknown lithological areas within the engineering site that are not covered by the borehole data, an empirical conversion formula for transverse and longitudinal wave velocities is used to convert the P-wave velocity corresponding to the three-dimensional absolute P-wave velocity model into S-wave velocity.

[0006] Optionally, acquiring the high-frequency surface wave dispersion data of the active source specifically includes: High-density linear flow arrays were deployed in the target areas of interest at the engineering site to conduct artificially excited active source surface wave exploration. After correcting the active source earthquake time, the vertical component waveforms were arranged and organized according to the epicentral distance. The Rayleigh wave phase velocity dispersion curve is extracted from the processed waveform data using the frequency-Bessel transform method. The Rayleigh wave phase velocity dispersion curve after quality screening is used as the high-frequency surface wave dispersion data of the active source.

[0007] Optionally, acquiring passive source surface wave dispersion data specifically includes: Continuous background noise observation is conducted based on a dense array of stations deployed in the engineering site area. The continuous noise data collected by each station is preprocessed, including instrument response removal, trend removal, mean removal, normalization, and bandpass filtering. Based on the preprocessed continuous noise data, the noise cross-correlation function between the same components between every two stations is calculated, including vertical-vertical components and tangential-tangential components. The noise cross-correlation functions of multiple time periods are superimposed to extract the empirical Green's function. The frequency-Bessel transform method is used to extract the dispersion curves of Rayleigh waves and Love waves from the empirical Green's function. After quality control and screening of the extracted dispersion curves, the dispersion data of the passive source surface wave is obtained.

[0008] Optionally, obtaining the individual HVSR curves of each station in the engineering site area specifically includes: Based on the three-component noise data collected by a single station, the spectral ratio of the horizontal component to the vertical component of a single station is calculated to obtain the HVSR curve of the corresponding station. The peak frequency of the single HVSR curve corresponds to the S-wave resonance frequency of the site strata, and is used to provide constraints on the thickness of the sedimentary layer and the burial depth of the bedrock interface for joint inversion.

[0009] Optionally, the integration of active source high-frequency surface wave dispersion data, passive source surface wave dispersion data, and single-station HVSR curves for joint inversion yields a three-dimensional refined S-wave velocity structure, specifically including: A joint inversion objective function is constructed. During the inversion solution process, the velocity structure of the ultra-shallow layer is constrained by the active source high-frequency surface wave dispersion data, the velocity structure of the mid-deep layer is constrained by the passive source surface wave dispersion data, the sedimentary layer thickness and bedrock interface burial depth parameters are constrained by the single HVSR curve, and the range of velocity parameter values ​​is constrained by the three-dimensional absolute P-wave velocity model. The three-dimensional fine S-wave velocity structure is obtained by iterative solution.

[0010] Optionally, the joint inversion is solved using a linear iterative inversion method. The inversion parameters are optimized through multiple iterations until the inversion results meet the fitting accuracy requirements, thereby obtaining the three-dimensional fine S-wave velocity structure.

[0011] Optionally, after outputting the final three-dimensional fine velocity structure model, the method further includes: The output three-dimensional fine velocity structure model is compared with existing geological data and borehole measurement data in the engineering site area to verify the reliability of the three-dimensional fine velocity structure model.

[0012] Optionally, the engineering site area is a hydropower or nuclear power engineering site area with a complex geological background of wide valleys and thick overburden layers. The output three-dimensional fine velocity structure model is used for seismic safety evaluation and seismic design of the engineering site.

[0013] Secondly, this invention also provides a system for constructing a three-dimensional fine velocity structure model of a complex engineering site, comprising: The construction module is used to acquire borehole data from the engineering site area and construct a one-dimensional initial velocity structure model based on the borehole data, including P-wave velocity, S-wave velocity, and P-wave / S-wave velocity ratio. The conversion module is used to conduct active source volume wave exploration based on a dense array of stations deployed in the engineering site area, obtain a three-dimensional absolute P-wave velocity model through volume wave first arrival travel time tomography inversion, and convert the three-dimensional absolute P-wave velocity model into an initial S-wave velocity model. The acquisition module is used to acquire active source high-frequency surface wave dispersion data, passive source surface wave dispersion data, and the single HVSR curves of each station in the engineering site area. The output module is used to take the one-dimensional initial velocity structure model as a reference, the initial S-wave velocity model as the initial inversion model, integrate the active source high-frequency surface wave dispersion data, the passive source surface wave dispersion data and the single HVSR curve for joint inversion, obtain the three-dimensional fine S-wave velocity structure, and output the final three-dimensional fine velocity structure model in combination with the three-dimensional absolute P-wave velocity model.

[0014] The method provided by this invention has the following beneficial effects: This method constructs a one-dimensional initial velocity structure model using borehole data from the engineering site area, providing a reference benchmark that conforms to the actual geological conditions of the site for the entire modeling process and preventing subsequent inversion results from deviating from regional geological patterns. A three-dimensional absolute P-wave velocity model obtained through active source body wave exploration provides stable absolute velocity calibration for subsequent joint inversion, narrowing the reasonable range of inversion parameter values ​​and effectively reducing the common problem of multiple solutions in the joint inversion of surface waves and HVSR. By integrating active source high-frequency surface wave dispersion data, passive source surface wave dispersion data, and single-station HVSR curves for joint inversion, the complementary advantages of different data types in velocity structure resolution can be fully utilized, achieving targeted constraints across the entire depth range from ultra-shallow to mid-deep, filling depth coverage blind spots of single data, and accurately identifying strong impedance interfaces of core engineering concern such as bedrock interfaces and sedimentary layer thickness. This significantly improves the overall accuracy and reliability of the velocity structure model, resulting in a refined three-dimensional velocity structure model that is suitable for the application requirements of seismic safety assessment and seismic design at major engineering sites. Attached Figure Description

[0015] Figure 1 A schematic diagram of the method for constructing a three-dimensional fine velocity structure model of a complex engineering site area provided in the embodiments of this application; Figure 2 This is a schematic diagram illustrating the multi-source data depth coverage and constraint effects provided in the embodiments of this application. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this application clearer, specific embodiments of this application will be described in further detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining this application and not for limiting it. It should also be noted that, for ease of description, only the parts relevant to this application are shown in the drawings, not all of them. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0017] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0018] This method for constructing a detailed 3D velocity structure model for complex engineering sites is primarily applied to hydropower and nuclear power plant sites with complex geological structures characterized by wide valleys and thick overburden layers. The output detailed 3D velocity structure model can be directly used for seismic safety assessment and seismic design of engineering sites. In earthquake prevention and disaster reduction for major engineering sites, the rationality of the site's ground motion input determines the safety of seismic design and the economy of engineering construction. The underground medium velocity structure model is the foundation for calculating site ground motion parameters. Especially for sites with complex geological conditions such as wide valleys and thick overburden layers, the underground strata vary greatly, and the overburden thickness is unevenly distributed, placing extremely high engineering requirements on the coverage, depth resolution, and interface recognition accuracy of the velocity structure model. Currently, in velocity structure modeling techniques for engineering sites, single methods such as borehole testing, active source exploration, and passive source imaging all have certain limitations in practical applications of complex sites. They cannot simultaneously consider the model's coverage, full-depth resolution, result reliability, and engineering implementation cost, and cannot fully meet the rigid requirements of seismic design for major engineering projects. This method integrates multi-source exploration data and combines the complementary advantages of different data types in terms of velocity structure resolution to achieve three-dimensional fine velocity structure modeling of complex engineering site areas from ultra-shallow to deep strata, providing reliable basic data support for the seismic safety assessment of engineering sites.

[0019] See Figures 1 to 2 This invention provides a method for constructing a three-dimensional fine velocity structure model of a complex engineering site, including the following steps: S1. Obtain borehole data from the engineering site area and construct a one-dimensional initial velocity structure model based on the borehole data, including P-wave velocity, S-wave velocity, and P-wave / S-wave velocity ratio.

[0020] The implementation of this method first requires acquiring borehole data from the engineering site area. Based on the borehole data, a one-dimensional initial velocity structure model is constructed, including P-wave velocity, S-wave velocity, and the P-wave / S-wave velocity ratio. Borehole data, as the most direct in-situ geological and geophysical test data of the engineering site area, can most accurately reflect the in-situ physical and mechanical properties of the strata. It is the most reliable geological reference for all inversion work, effectively avoiding systematic deviations in subsequent inversion results that do not match the actual geological conditions. Borehole data is typically obtained through in-situ wave velocity testing and borehole core sampling laboratory tests during the early stages of engineering exploration. The data includes lithology of the strata at different borehole depths, measured values ​​of P-wave velocity and S-wave velocity, and the P-wave / S-wave velocity ratio at different depths calculated based on the measured data. It also includes key geological information such as stratigraphic layering, overburden thickness, and bedrock surface burial depth revealed by the borehole.

[0021] After data verification, the wave velocity data from different boreholes were statistically analyzed in layers, taking into account the overall geological background of the site area and the stratigraphic distribution patterns revealed by the boreholes. This determined the reasonable ranges for P-wave velocity, S-wave velocity, and P-wave / S-wave velocity ratios for each stratum in the site area, and based on this, a one-dimensional initial velocity structure model was constructed. In constructing the one-dimensional initial velocity structure model, the stability and convergence requirements of subsequent inversion work were carefully considered. The reliability of the results of subsequent multi-source data joint inversion largely depends on the fit between the one-dimensional initial velocity structure model and the actual geological conditions of the site. If the one-dimensional initial velocity structure model has excessive deviation, the inversion process may converge to a local optimum, resulting in results deviating from actual geological conditions. The constructed one-dimensional initial velocity structure model reflects the overall wave velocity variation patterns and vertical stratification characteristics of the strata in the engineering site area, providing a reference benchmark that conforms to the actual geological conditions of the site for subsequent full-process inversion calculations. This avoids results that deviate from the true geological conditions of the site during subsequent inversion processes, and also provides a basic geological basis for setting the value range of subsequent inversion parameters and verifying the inversion results.

[0022] S2. Active source volume wave exploration is carried out based on a dense array of stations deployed in the engineering site area. The three-dimensional absolute P-wave velocity model is obtained by inversion of the first arrival travel time tomography of the volume wave, and the three-dimensional absolute P-wave velocity model is converted into the initial S-wave velocity model.

[0023] In some implementations, the three-dimensional absolute P-wave velocity model is converted into an initial S-wave velocity model, specifically including: For areas within the engineering site covered by borehole data, the P-wave velocity corresponding to the three-dimensional absolute P-wave velocity model is converted into S-wave velocity using the measured P-wave velocity to S-wave velocity ratio from the borehole data. For unknown lithological areas within the engineering site where no borehole data is available, empirical conversion formulas for transverse and longitudinal wave velocities are used to convert the P-wave velocity corresponding to the three-dimensional absolute P-wave velocity model into S-wave velocity.

[0024] Based on the construction of the one-dimensional initial velocity structure model, active source body wave exploration was conducted using a dense array of stations deployed in the engineering site area. A three-dimensional absolute P-wave (Primary Wave) velocity model was obtained through body wave first-arrival travel-time tomography inversion, and then converted into an initial S-wave (Secondary Wave) velocity model. This step focused on addressing the common problems of insufficient absolute velocity constraints and multiple solutions in the subsequent joint inversion of surface waves and HVSR curves. These problems are particularly prominent in complex sites with wide valleys and deep overburden. Relying solely on surface wave and HVSR data for inversion often results in overall velocity shifts and errors in identifying the bedrock interface depth, failing to meet the rigid accuracy requirements of seismic design for major engineering projects. Therefore, the absolute P-wave velocity obtained through active source body wave exploration provides stable absolute velocity calibration for the entire inversion process. The implementation of active source body wave exploration is based on the completed one-dimensional initial velocity structure model. The array layout is designed in conjunction with the topographic conditions, exploration range, and depth requirements of the engineering site area. This ensures that the array coverage completely encompasses the core area of ​​the engineering site, and the station spacing is matched to the resolution requirements of the target exploration depth. This allows the acquired waveform data to fully reflect the wavefield characteristics of the shallow to mid-deep strata in the site area. The excitation method of the active source can be selected according to the site conditions, using either artificial excitation or controlled-source vehicle excitation. This ensures that the excitation energy is effectively transferred to the strata at the target exploration depth, while also ensuring that the parameters of each excitation are stable and controllable, guaranteeing the consistency and signal-to-noise ratio of the acquired data.

[0025] After acquiring waveform data from the active source body wave exploration, the waveform data corresponding to each excitation is preprocessed to correct the active source's origin time and extract the effective waveform data for the corresponding excitation period. Then, the P-wave first arrival travel time data for each shot is extracted. Subsequently, using a one-dimensional initial velocity structure model as the inversion basis, body wave first arrival travel time tomography is employed for inversion calculations to obtain a three-dimensional absolute P-wave velocity model for the engineering site area. This three-dimensional absolute P-wave velocity model clearly reflects the three-dimensional spatial distribution characteristics of the P-wave velocity in the site area, accurately describes the spatial undulations of the stratigraphic interfaces, and provides absolute velocity calibration for subsequent joint inversions, narrowing the reasonable range of inversion parameters, reducing the ambiguity of inversion results, and compensating for the inherent limitation of lacking absolute velocity constraints in the joint inversion of surface waves and HVSR curves.

[0026] After obtaining the three-dimensional absolute P-wave velocity model, and considering the borehole data coverage of the engineering site area, a differentiated conversion rule was adopted to convert the three-dimensional absolute P-wave velocity model into an initial S-wave velocity model. During the initial S-wave velocity model conversion process, taking into account the uneven distribution of borehole data in the engineering site area, a differentiated conversion rule was adopted for different regions to ensure the accuracy and rationality of the initial S-wave velocity model across the entire site. For areas within the engineering site area covered by borehole data, the measured P-wave velocity to S-wave velocity ratio from the corresponding borehole data was directly used to convert the P-wave velocity at the corresponding location in the three-dimensional absolute P-wave velocity model into S-wave velocity. This ensured that the converted S-wave velocity highly matched the in-situ measured data of the site area, closely reflecting the actual physical characteristics of the regional strata and avoiding systematic biases caused by empirical conversion. For unknown lithological areas within the engineering site where no borehole data is available, an industry-standard empirical conversion formula for P-wave and S-wave velocities is used to convert the P-wave velocities at corresponding locations in the 3D absolute P-wave velocity model into S-wave velocities. This ensures that the S-wave velocity values ​​in borehole-free areas conform to the general variation patterns of stratigraphic lithology, filling the parameter gaps in areas without measured data. The converted initial S-wave velocity model provides reasonable initial parameters for subsequent joint inversion, defining a reasonable range for S-wave velocity values, and allowing for optimization and adjustment of the S-wave velocity parameters within this reasonable range during subsequent joint inversion.

[0027] S3. Obtain the high-frequency surface wave dispersion data of the active source, the surface wave dispersion data of the passive source, and the single HVSR curves of each station in the engineering site area.

[0028] In some implementations, acquiring high-frequency surface wave dispersion data from an active source specifically includes: High-density linear flow arrays were deployed in the target areas of interest at the engineering site to conduct artificially excited active source surface wave exploration. After correcting the active source earthquake time, the vertical component waveforms were arranged and organized according to the epicentral distance. The Rayleigh wave phase velocity dispersion curve was extracted from the processed waveform data using the frequency-Bessel transform method. The Rayleigh wave phase velocity dispersion curve after quality screening was used as the high-frequency surface wave dispersion data of the active source.

[0029] In some implementations, acquiring passive source surface wave dispersion data specifically includes: Continuous background noise observation is conducted using a dense array of monitoring stations deployed in the engineering site area. The continuous noise data collected by each station is preprocessed, including instrument response removal, trend removal, mean removal, normalization, and bandpass filtering. Based on the preprocessed continuous noise data, the noise cross-correlation function between the same components between every two stations is calculated, including vertical-vertical components and tangential-tangential components. The noise cross-correlation functions of multiple time periods are superimposed to extract the empirical Green's function. The frequency-Bessel transform method is used to extract the dispersion curves of Rayleigh waves and Love waves from the empirical Green's function. After quality control and screening of the extracted dispersion curves, the dispersion data of passive source surface waves are obtained.

[0030] In some implementations, the individual HVSR curves of each station in the engineering site area are obtained, specifically including: Based on the three-component noise data collected by a single station, the spectral ratio of the horizontal component to the vertical component of a single station is calculated to obtain the HVSR curve of the corresponding station. The peak frequency of a single HVSR curve corresponds to the S-wave resonance frequency of the basic strata at the site, and is used to provide constraints on the thickness of the sedimentary layer and the burial depth of the bedrock interface for joint inversion.

[0031] After completing the initial S-wave velocity model construction, multi-source constraint data were simultaneously acquired and processed to obtain active source high-frequency surface wave dispersion data, passive source surface wave dispersion data, and individual HVSR curves for each station in the engineering site area. This method considers the resolution requirements of complex engineering sites across the entire depth range from ultra-shallow to mid-deep. Single-band surface wave data often only covers a limited depth range, and the insufficient resolution of ultra-shallow strata and limited coverage in mid-deep strata are particularly prominent in sites with deep overburden. Therefore, by combining different types and frequency bands of surface wave data with HVSR curves, targeted constraints on strata at different depths are achieved, filling the coverage blind spots of single data and providing full-dimensional parameter constraints for subsequent joint inversion.

[0032] For ultra-shallow strata, a core concern in the seismic design of major engineering projects, passively sourced surface wave data often lacks high-frequency information and has insufficient resolution for strata within a depth of tens of meters near the surface. This makes it impossible to accurately describe the strata layering and wave velocity variations in ultra-shallow strata, failing to meet the accuracy requirements of strata parameters in engineering foundation design. Therefore, high-density linear active-source surface wave exploration is employed to acquire high-frequency surface wave dispersion data, providing strong constraints for ultra-shallow strata. In specific implementation, high-density linear mobile arrays are deployed in the target areas of the engineering site, including the core construction area of ​​engineering structures, key sections with drastic strata changes, and critical areas with complex geological conditions. The station spacing is set according to the resolution requirements of ultra-shallow exploration to ensure effective acquisition of complete high-frequency surface wave signals. Artificially excited active-source surface wave exploration is then conducted. After data acquisition, the active source excitation time is corrected, and the vertical component waveforms are arranged and organized according to the epicentral distance. Anomalous and invalid traces are removed, and the influence of environmental interference on the waveform data is eliminated to ensure the signal-to-noise ratio and consistency of the waveform data. Subsequently, the frequency-Bessel transform method was used to extract Rayleigh wave phase velocity dispersion curves from the processed waveform data. The extracted dispersion curves were then subjected to quality screening to remove invalid curves with abnormal dispersion trends or non-concentrated energy. The Rayleigh wave phase velocity dispersion curves after quality screening were used as active source high-frequency surface wave dispersion data. This type of data contains rich high-frequency information, which can effectively cover ultra-shallow strata, make up for the natural deficiencies of passive source data in the high-frequency band, and greatly improve the inversion resolution of ultra-shallow strata, so that the final model can accurately reflect the near-surface strata characteristics that are of core concern in engineering construction.

[0033] To address the need for extensive coverage of deep strata within the engineering site area, active source exploration is limited by excitation energy, implementation costs, and site conditions, making it difficult to achieve continuous, large-scale planar coverage across the entire site. Therefore, continuous background noise observation is employed to acquire low-frequency passive source surface wave dispersion data, achieving uniform constraint of deep strata across the entire site. In practice, continuous background noise observation is conducted using a dense array of stations deployed across the engineering site area. The observation duration (generally several weeks to several months) is set based on the energy intensity of the low-frequency components in the background noise of the site area, ensuring that effective low-frequency surface wave signals matching the target exploration depth can be acquired through cross-correlation superposition. The continuous noise data acquired by each station undergoes preprocessing, including instrument response removal, trend removal, mean removal, normalization, and bandpass filtering. This process systematically eliminates the impact of instrument system errors, long-period trend terms, and sudden environmental interference on data quality, providing high-quality foundational data for subsequent data processing. Based on the preprocessed continuous noise data, the noise cross-correlation function between the same components of every two stations is calculated, specifically including vertical-vertical and tangential-tangential components. The noise cross-correlation functions of multiple time periods are superimposed to effectively improve the signal-to-noise ratio of the effective signal, suppress random interference, and extract the empirical Green's function. Then, the frequency-Bessel transform method is used to extract the dispersion curves of Rayleigh and Love waves from the empirical Green's function. The extracted dispersion curves undergo quality control and screening, eliminating invalid data with insufficient signal-to-noise ratio, unclear dispersion characteristics, or abnormal trends. Finally, passive source surface wave dispersion data is obtained. This type of data can effectively cover the deep strata of the site area, achieving continuous surface coverage of the entire site. It perfectly complements the active source high-frequency surface wave dispersion data in terms of depth, completely covering all target strata from ultra-shallow to medium-deep, eliminating depth coverage blind spots of single data.

[0034] Based on the acquisition and processing of surface wave dispersion data, single-station HVSR curves are simultaneously acquired. This method addresses the insufficient ability of surface wave inversion to identify strong impedance interfaces such as sedimentary layer thickness and bedrock interfaces. These interfaces are a core focus of seismic safety assessment for engineering sites, and deviations in interface depth identification affect the accuracy of ground motion parameter calculations, thereby impacting the safety and economy of seismic design. HVSR curves are highly sensitive to these strong impedance interfaces, providing accurate and independent constraints on interface depth. In practice, based on the three-component noise data acquired by a single station, the spectral ratio of the horizontal to vertical components is calculated to obtain the corresponding single-station HVSR curve. The peak frequency of the single-station HVSR curve corresponds to the fundamental S-wave resonance frequency of the site strata, directly reflecting the resonance characteristics of the sedimentary layer below the station. This provides constraints on sedimentary layer thickness and bedrock interface depth for joint inversion, compensating for the inherent limitations of surface wave dispersion data in identifying strong impedance interfaces, further narrowing the range of inversion parameter values, reducing the ambiguity of inversion results, and making the final velocity structure model more closely match the actual geological conditions of the site. See also: The depth coverage and constraints of multi-source data Figure 2 The ultra-shallow layer is the topmost stratum near the surface, directly impacting the engineering foundation. It has the smallest vertical extent, and its geological parameters have the most direct influence on engineering design. The intermediate-shallow layer is the main section of the shallow overburden below the ultra-shallow layer, with a larger vertical extent. It is a core component of the site's overburden, connecting the near-surface and deep strata. The intermediate-deep layer is the deep section of the overburden extending from the intermediate-shallow layer to the bedrock surface. It is the main extension of the deep overburden in wide valley sites, with the largest vertical scale, making it difficult to effectively cover using conventional shallow exploration. The deep layer is the complete bedrock distribution area below the bedrock surface, serving as the bottom benchmark layer of the site's stratigraphic structure and providing stable geological constraints for velocity modeling across the entire site. Among them, borehole data is uniformly distributed along the horizontal range of the site, generally reaching the bedrock interface, that is, mainly affecting the ultra-shallow, intermediate-shallow, and intermediate-deep layers; the three-dimensional absolute P-wave velocity model mainly affects the entire depth range from ultra-shallow to deep layers, providing full-range absolute velocity calibration for subsequent joint inversion and narrowing the reasonable range of inversion parameters; active source high-frequency surface wave dispersion data mainly affects the ultra-shallow layer, providing high-resolution constraints for the ultra-shallow strata of core engineering interest; passive source surface wave dispersion data mainly affects the intermediate-shallow, intermediate-deep, and deep layers, providing uniform surface constraints for the intermediate-deep strata of the site; a single HVSR curve can provide accurate constraints on the spatial distribution of sedimentary layer thickness (ultra-shallow, intermediate-shallow, and intermediate-deep layers together constitute a complete sedimentary layer) and bedrock interface, and can also provide certain constraints on S-wave velocity structure.

[0035] S4. Using the one-dimensional initial velocity structure model as a reference and the initial S-wave velocity model as the initial inversion model, the active source high-frequency surface wave dispersion data, passive source surface wave dispersion data and single-station HVSR curves are integrated for joint inversion to obtain the three-dimensional fine S-wave velocity structure. Combined with the three-dimensional absolute P-wave velocity model, the final three-dimensional fine velocity structure model is output.

[0036] In some implementations, active source high-frequency surface wave dispersion data, passive source surface wave dispersion data, and single-station HVSR curves are integrated for joint inversion to obtain a three-dimensional fine S-wave velocity structure, specifically including: A joint inversion objective function is constructed. During the inversion solution process, the velocity structure of the ultra-shallow layer is constrained by the high-frequency surface wave dispersion data of the active source, the velocity structure of the mid-deep layer is constrained by the surface wave dispersion data of the passive source, the sedimentary layer thickness and the burial depth parameters of the bedrock interface are constrained by the single HVSR curve, and the range of velocity parameter values ​​is constrained by the three-dimensional absolute P-wave velocity model. The three-dimensional fine S-wave velocity structure is obtained by iterative solution.

[0037] In some implementations, the joint inversion is solved using a linear iterative inversion method. The inversion parameters are optimized through multiple iterations until the inversion results meet the fitting accuracy requirements, thereby obtaining a three-dimensional fine S-wave velocity structure.

[0038] After completing the construction of the one-dimensional initial velocity structure model, the three-dimensional absolute P-wave velocity model, and the initial S-wave velocity model, as well as acquiring active source high-frequency surface wave dispersion data, passive source surface wave dispersion data, and single-station HVSR curves, multi-source data joint inversion was performed to finally obtain a three-dimensional fine velocity structure model of the engineering site area. Considering the inherent limitations of single-type data inversion, although single surface wave dispersion data inversion can obtain the shear wave velocity structure of the subsurface medium, it is limited by the data frequency band coverage and cannot simultaneously take into account the resolution of ultra-shallow and intermediate-deep subsurfaces. Furthermore, it lacks independent constraints on absolute velocity and key geological interfaces, resulting in highly ambiguous inversion results that are difficult to meet the modeling accuracy requirements of complex engineering sites. Therefore, by jointly inverting multi-source data, the advantages of different types of data in velocity structure resolution are fully combined to achieve accurate constraints across the entire depth range, fundamentally improving the reliability and accuracy of the inversion results.

[0039] Before conducting the full-site 3D joint inversion, the acquired active-source high-frequency surface wave dispersion data is first processed using one-dimensional inversion. Based on the quality-screened active-source high-frequency surface wave dispersion data, the one-dimensional S-wave velocity structure of the ultra-shallow layer in the target area of ​​interest is obtained through inversion. This inverted one-dimensional S-wave velocity structure can accurately reflect the vertical variation characteristics of the shear wave velocity in the ultra-shallow strata of the target area, providing prior velocity constraints for the ultra-shallow layer in the subsequent full-site 3D joint inversion. This further narrows the reasonable range of values ​​for the ultra-shallow strata inversion parameters, avoids unreasonable offsets in near-surface velocity parameters during the joint inversion process, and ensures the inversion accuracy of the ultra-shallow strata of core concern for the engineering construction. Furthermore, before conducting the full-site 3D joint inversion, the vertical grid of the inversion model can be differentiated based on the effective depth coverage and inherent resolution of the multi-source data. For the ultra-shallow subsurface regions effectively covered by active-source high-frequency surface wave dispersion data, a smaller vertical grid step size is adopted to match the vertical resolution of the high-frequency surface wave data, avoiding the loss of ultra-shallow stratigraphic details caused by excessively large grid scales. For the mid-deep and deep subsurface regions effectively covered by passive-source surface wave dispersion data, a grid step size matching the data wavelength is adopted to ensure the stability of the inversion calculation. For the bedrock strong impedance interface locations locked by a single HVSR curve, a forced interface for grid partitioning is set to ensure that the undulation morphology of the bedrock interface can be accurately restored in the inversion model, avoiding stratigraphic interface identification bias caused by cross-interface grid partitioning. After completing the preparation of ultra-shallow prior constraint parameters, a joint inversion objective function is constructed. The construction of the objective function fully incorporates the fitting residual terms of various types of data, while also integrating the corresponding constraint conditions of each type of data, ensuring that the advantages of various types of data can be fully utilized during the inversion process, avoiding the fitting bias of a single data point from dominating the inversion results, and ensuring the stability of the inversion process and the rationality of the results.

[0040] During the inversion process, targeted constraints are applied based on the stratigraphic characteristics and inversion requirements of different depth segments. Using active source high-frequency surface wave dispersion data and the inverted one-dimensional S-wave velocity structure of the ultra-shallow layer, the velocity structure of the ultra-shallow layer is specifically constrained during the inversion process, ensuring the inversion resolution and accuracy of the near-surface strata and meeting the rigid requirements of near-surface strata parameters in engineering seismic design. Passive source surface wave dispersion data is used to specifically constrain the velocity structure of the mid-deep layers during the inversion process, achieving uniform coverage and stable constraints of the mid-deep strata across the entire site, filling the mid-deep depth blind spots that are difficult to cover with active source data. The system employs a single HVSR curve to constrain the sedimentary layer thickness and bedrock interface depth parameters during the inversion process. This accurately identifies the spatial distribution of the high-impedance interface, a key focus of the project, compensating for the insufficient ability of surface wave data to identify stratigraphic interfaces and preventing systematic biases in bedrock surface depth inversion. Furthermore, a pre-acquired three-dimensional absolute P-wave velocity model constrains the velocity parameter range throughout the inversion process, providing stable absolute velocity calibration and reducing the ambiguity of the inversion results from the outset. This prevents overall velocity shifts and ensures the inversion results consistently reflect the actual geological conditions of the site area. Additionally, during the inversion constraint process, the constraint weights for different data sources can be differentiated based on the spatial distribution differences in geological conditions within the engineering site area. For key areas of concern within the site with complex geological conditions, dramatic variations in overburden thickness, and large stratigraphic undulations, including the core riverbed area, areas adjacent to fault structures, and areas planned for building construction, the constraint weights of active source high-frequency surface wave dispersion data and single-station HVSR curves are appropriately increased to further enhance the inversion resolution of ultra-shallow strata and the identification accuracy of bedrock interfaces. For areas within the site with stable geological conditions and uniform stratigraphic distribution, including bedrock outcrops on riverbank slopes and areas with gentle stratigraphic changes, the constraint weights of the three-dimensional absolute P-wave velocity model and passive source surface wave dispersion data are appropriately increased to ensure the convergence efficiency of the inversion process and the overall stability of the results across the entire site.

[0041] The joint inversion employs a linear iterative inversion method. This method was chosen because of the computational efficiency and result stability requirements of engineering applications. An initial S-wave velocity model closely matching the actual geological conditions of the site was constructed beforehand, along with multi-dimensional constraints. The linear iterative inversion method can quickly converge to a reasonable optimal solution based on the existing initial S-wave velocity model and constraints, significantly improving the efficiency of the inversion calculation and meeting the time-sensitive requirements of engineering site modeling. In the specific implementation, the initial S-wave velocity model serves as the starting point for iteration. Combined with the constructed joint inversion objective function and multi-dimensional constraints, the inversion parameters are optimized through multiple iterations. After each iteration, the fitting residuals of various data are calculated, and it is determined whether the residuals meet the preset fitting accuracy requirements. If not, the velocity model parameters are updated based on the residual results, and the next iteration continues until the inversion results meet the fitting accuracy requirements, at which point the iteration terminates, ultimately yielding the three-dimensional refined S-wave velocity structure of the engineering site area. The inverted three-dimensional fine S-wave velocity structure is spatially matched and integrated with the previously acquired three-dimensional absolute P-wave velocity model to form a complete three-dimensional fine velocity structure model of the engineering site area from the ultra-shallow to the deep strata. This model contains three-dimensional spatial distribution information of both P-wave and S-wave velocities, and can fully reflect the velocity variation characteristics of the underground medium and the undulation morphology of the strata interface in the site area. It can directly provide reliable basic data support for the subsequent seismic safety assessment and seismic design of the engineering site.

[0042] In some implementations, after outputting the final three-dimensional fine velocity structure model, the method further includes: The output three-dimensional fine velocity structure model was compared with the existing geological data and borehole measurement data of the engineering site area to verify the reliability of the three-dimensional fine velocity structure model.

[0043] After obtaining the final three-dimensional fine velocity structure model, the reliability of the model can be verified. In the specific implementation process, the existing geological data and borehole measurement data of the engineering site area are first organized. The geological data includes the regional geological survey report of the site area, engineering geological mapping results, stratigraphic zoning and fault structure distribution data, etc., which can comprehensively reflect the overall geological background and stratigraphic spatial distribution law of the site area. The borehole measurement data includes the in-situ wave velocity test results corresponding to different depths of each borehole, stratigraphic layering information, measured values ​​of bedrock surface burial depth, etc., which are the most direct in-situ measurement data for verifying the accuracy of the model. During the verification process, the spatial distribution characteristics and undulating morphology of stratigraphic interfaces reflected by the 3D fine velocity structure model were first compared with the geological laws revealed by the geological data of the site area to determine whether the model conformed to the overall geological background of the site area and whether there were any abnormal results that contradicted the regional geological laws. Then, the calculated values ​​of vertical velocity profiles, stratigraphic layer depths, and bedrock surface burial depths at the corresponding borehole locations in the model were compared point-to-point with the measured borehole data to analyze the deviation range between the calculated and measured values ​​and verify the vertical resolution and accuracy of the model. Through this dual verification of overall geological law comparison and point-to-point verification with measured borehole data, the reliability of the 3D fine velocity structure model can be confirmed, ensuring that the model results can accurately reflect the actual underground medium characteristics of the site area and meet the accuracy requirements of major engineering applications.

[0044] In some implementations, the engineering site area is a hydropower or nuclear power engineering site area with a complex geological background of wide valleys and thick overburden. The output three-dimensional fine velocity structure model is used for seismic safety evaluation and seismic design of the engineering site.

[0045] This method is primarily applicable to hydropower and nuclear power plant sites with complex geological structures characterized by wide valleys and thick overburden. These sites often exhibit significant topographic relief, uneven spatial distribution of strata, dramatic variations in overburden thickness, and complex bedrock surface undulations. They demand extremely high standards for the coverage, full-depth resolution, and interface identification accuracy of the subsurface velocity structure model. Conventional single-source exploration methods are insufficient to meet these modeling requirements. This method, through the fusion of multi-source data, perfectly adapts to the modeling needs of such complex sites. The output 3D high-precision velocity structure model can be directly used for seismic safety assessment of engineering sites, calculation of site ground motion parameters, and seismic design, providing reliable basic data support for earthquake prevention and disaster reduction in major projects. This method has been validated through practical engineering applications, successfully applied to the 3D high-precision velocity structure modeling of a hydropower project site with a wide valley and thick overburden. This site possesses typical complex engineering geological conditions, including a wide valley, a thick and highly unevenly distributed riverbed overburden, dramatic variations in strata on both banks, and complex regional geological structures. Conventional exploration methods struggle to achieve high-precision modeling across the entire site and depth. In engineering applications, following the implementation process of this method, existing borehole data from the site area are first collected to construct a one-dimensional initial velocity structure model. Then, active source volume wave exploration is conducted based on a dense array of stations deployed in the site area to obtain a three-dimensional absolute P-wave velocity model, which is then converted to obtain an initial S-wave velocity model. Simultaneously, active source high-frequency surface wave exploration, continuous passive source background noise observation, and HVSR curve calculation are performed to obtain multi-source constrained data. Afterwards, multi-source data joint inversion is performed to obtain a three-dimensional refined velocity structure model of the site area, and finally, the model reliability is verified. Application results show that the three-dimensional refined velocity structure model constructed by this method can clearly describe the spatial distribution characteristics of velocity from ultra-shallow to deep strata in the site area, accurately describe the undulation morphology of the bedrock surface and the spatial variation law of the overburden thickness. The model results show high agreement with the borehole measured data, and the resolution of the ultra-shallow strata can fully meet the needs of engineering foundation design and seismic design, effectively supporting the subsequent seismic safety assessment of the site area, and fully verifying the engineering practicality and reliability of this method.

[0046] The design concept of this method remains unchanged regardless of adjustments to data processing methods, inversion algorithms, and exploration implementation methods during the specific implementation process. In practical engineering applications, suitable alternative implementation schemes can be selected based on the site conditions, data quality, and engineering requirements, all achieving the same technical effect. For the surface wave dispersion curve extraction stage, the preferred implementation scheme of this method is the frequency-Bessel transform method. This method is adaptable to observation data from dense arrays, effectively extracting multi-order surface wave dispersion curves, and its computational efficiency and extraction accuracy meet the requirements of engineering applications. In practical applications, other suitable dispersion curve extraction methods can also be selected based on the array layout and the signal-to-noise ratio of the data, including the dual-array method, the improved frequency-Bessel transform method, and the clustering analysis method. All these methods can effectively extract surface wave dispersion curves from waveform data, providing basic data for subsequent inversion, and are all alternative implementation schemes of this method. Regarding the solution algorithm for joint inversion, the preferred implementation of this method is the linear iterative inversion method. This method has high computational efficiency and fast convergence speed, and can perfectly meet the time requirements of engineering applications. In practical applications, if the geological conditions of the site area are extremely complex and the constraint ability of the initial inversion model is limited, other suitable inversion solution algorithms can also be selected, including nonlinear global search algorithms such as genetic algorithms, simulated annealing algorithms, and neighborhood algorithms, as well as Bayesian probabilistic inversion methods such as Markov chain Monte Carlo. All of these algorithms can achieve joint inversion solution of multi-source data and obtain the three-dimensional fine S-wave velocity structure of the site area, and are all alternative implementation schemes of this method. Regarding the excitation methods for active source exploration, the preferred excitation methods for active source volume wave exploration and active source surface wave exploration in this method are controlled seismic source vehicle excitation and manual hammer excitation, respectively, which can be adapted to the site conditions of most engineering sites. In practical applications, other suitable excitation methods, including blasting excitation and air gun excitation, can also be selected according to the topographic conditions of the site area, the target exploration depth, and on-site construction restrictions. As long as the excitation energy is stable, the signal can be effectively transmitted to the target exploration depth, and high-quality waveform data can be obtained, the same exploration effect can be achieved, which are alternative implementation schemes of this method.

[0047] In summary, this method addresses the core engineering need for detailed 3D velocity structure modeling at major engineering sites with complex geological structures characterized by wide valleys and deep overburden layers. It constructs a comprehensive modeling scheme integrating multi-source exploration data. The core idea of ​​this method is to address the inherent limitations of existing single exploration technologies in complex site applications by selectively fusing different types of exploration data to achieve complementary advantages. To address the common problems of insufficient absolute velocity constraints and multiple solutions in surface wave joint inversion, the method provides full-range calibration using 3D absolute P-wave velocities obtained from active source body wave exploration. To address the issue that passive source surface wave data alone cannot achieve full-depth resolution, the method combines active and passive source surface wave data to achieve full-depth coverage from ultra-shallow to mid-deep layers. To address the insufficient ability of surface wave inversion to identify key geological interfaces, the method provides accurate constraints using HVSR curves. The resulting detailed 3D velocity structure model is suitable for the seismic design requirements of major engineering projects and possesses strong engineering practicality and reliability.

[0048] This invention also provides a system for constructing a three-dimensional fine velocity structure model of a complex engineering site, including: The module is used to acquire borehole data from the engineering site area and construct a one-dimensional initial velocity structure model based on the borehole data, including P-wave velocity, S-wave velocity, and P-wave / S-wave velocity ratio. The conversion module is used for active source volume wave exploration based on a dense array of stations deployed in the engineering site area. It obtains a three-dimensional absolute P-wave velocity model through volume wave first arrival travel time tomography inversion and converts the three-dimensional absolute P-wave velocity model into an initial S-wave velocity model. The acquisition module is used to acquire active source high-frequency surface wave dispersion data, passive source surface wave dispersion data, and the single HVSR curves of each station in the engineering site area. The output module is used to take a one-dimensional initial velocity structure model as a reference, take an initial S-wave velocity model as the initial inversion model, integrate active source high-frequency surface wave dispersion data, passive source surface wave dispersion data and single HVSR curves for joint inversion, obtain a three-dimensional fine S-wave velocity structure, and output the final three-dimensional fine velocity structure model by combining the three-dimensional absolute P-wave velocity model.

[0049] The system embodiments provided in this invention have the same technical features as the method embodiments described above, and therefore can achieve the same technical effects, which will not be repeated here.

[0050] The above description is merely a preferred embodiment and the technical principles employed in this application. This application is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of this application.

Claims

1. A method for constructing a three-dimensional fine velocity structure model of a complex engineering site area, characterized in that, Includes the following steps: S1. Obtain borehole data from the engineering site area, and construct a one-dimensional initial velocity structure model based on the borehole data, including P-wave velocity, S-wave velocity, and P-wave / S-wave velocity ratio. S2. Active source volume wave exploration is carried out based on the dense array of stations deployed in the engineering site area. The three-dimensional absolute P-wave velocity model is obtained by inversion of the first arrival travel time tomography of the volume wave, and the three-dimensional absolute P-wave velocity model is converted into the initial S-wave velocity model. S3. Obtain the high-frequency surface wave dispersion data of the active source, the surface wave dispersion data of the passive source, and the single HVSR curve of each station in the engineering site area. S4. Using the one-dimensional initial velocity structure model as a reference and the initial S-wave velocity model as the inversion initial model, the active source high-frequency surface wave dispersion data, the passive source surface wave dispersion data and the single HVSR curve are integrated for joint inversion to obtain a three-dimensional fine S-wave velocity structure. The final three-dimensional fine velocity structure model is then output by combining the three-dimensional absolute P-wave velocity model. Specifically, converting the three-dimensional absolute P-wave velocity model into an initial S-wave velocity model includes: For areas within the engineering site covered by the borehole data, the P-wave velocity corresponding to the three-dimensional absolute P-wave velocity model is converted into S-wave velocity using the measured P-wave velocity to S-wave velocity ratio from the borehole data. For unknown lithological areas within the engineering site that are not covered by the borehole data, an empirical conversion formula for transverse and longitudinal wave velocities is used to convert the P-wave velocity corresponding to the three-dimensional absolute P-wave velocity model into S-wave velocity. The acquisition of active source high-frequency surface wave dispersion data specifically includes: High-density linear flow arrays were deployed in the target areas of interest at the engineering site to conduct artificially excited active source surface wave exploration. After correcting the active source earthquake time, the vertical component waveforms were arranged and organized according to the epicentral distance. The frequency-Bessel transform method is used to extract the Rayleigh wave phase velocity dispersion curve from the processed waveform data. The Rayleigh wave phase velocity dispersion curve after quality screening is used as the high-frequency surface wave dispersion data of the active source. Obtaining passive source surface wave dispersion data specifically includes: Continuous background noise observation is conducted based on a dense array of stations deployed in the engineering site area. The continuous noise data collected by each station is preprocessed, including instrument response removal, trend removal, mean removal, normalization, and bandpass filtering. Based on the preprocessed continuous noise data, the noise cross-correlation function between the same components between every two stations is calculated, including vertical-vertical components and tangential-tangential components. The noise cross-correlation functions of multiple time periods are superimposed to extract the empirical Green's function. The frequency-Bessel transform method is used to extract the dispersion curves of Rayleigh waves and Love waves from the empirical Green's function. After quality control and screening of the extracted dispersion curves, the dispersion data of the passive source surface wave is obtained. Obtain the individual HVSR curves for each station in the engineering site area, specifically including: Based on the three-component noise data collected by a single station, the spectral ratio of the horizontal component to the vertical component of a single station is calculated to obtain the HVSR curve of the corresponding station. The peak frequency of the single HVSR curve corresponds to the S-wave resonance frequency of the site strata, and is used to provide constraints on the thickness of the sedimentary layer and the burial depth of the bedrock interface for joint inversion.

2. The method of claim 1, wherein, The process of integrating the active source high-frequency surface wave dispersion data, the passive source surface wave dispersion data, and the single-station HVSR curve for joint inversion yields a three-dimensional refined S-wave velocity structure, specifically including: A joint inversion objective function is constructed. During the inversion solution process, the velocity structure of the ultra-shallow layer is constrained by the active source high-frequency surface wave dispersion data, the velocity structure of the mid-deep layer is constrained by the passive source surface wave dispersion data, the sedimentary layer thickness and bedrock interface burial depth parameters are constrained by the single HVSR curve, and the range of velocity parameter values ​​is constrained by the three-dimensional absolute P-wave velocity model. The three-dimensional fine S-wave velocity structure is obtained by iterative solution.

3. The method of claim 2, wherein, The joint inversion is solved using a linear iterative inversion method. The inversion parameters are optimized through multiple iterations until the inversion results meet the fitting accuracy requirements, thereby obtaining the three-dimensional fine S-wave velocity structure.

4. The method of claim 1, wherein, After outputting the final three-dimensional fine velocity structure model, the method further includes: The output three-dimensional fine velocity structure model is compared with the existing geological data and borehole measurement data of the engineering site area to verify the reliability of the three-dimensional fine velocity structure model.

5. The method of claim 1, wherein, The project site area is a complex geological structure area with wide river valleys and thick overburden layers, suitable for hydropower and nuclear power projects. The output three-dimensional fine velocity structure model is used for seismic safety evaluation and seismic design of the project site.

6. A system for constructing a three-dimensional fine velocity structure model of a complex engineering site area, characterized by, include: The construction module is used to acquire borehole data from the engineering site area and construct a one-dimensional initial velocity structure model based on the borehole data, including P-wave velocity, S-wave velocity, and P-wave / S-wave velocity ratio. The conversion module is used to conduct active source volume wave exploration based on a dense array of stations deployed in the engineering site area, obtain a three-dimensional absolute P-wave velocity model through volume wave first arrival travel time tomography inversion, and convert the three-dimensional absolute P-wave velocity model into an initial S-wave velocity model. The acquisition module is used to acquire active source high-frequency surface wave dispersion data, passive source surface wave dispersion data, and the single HVSR curves of each station in the engineering site area. The output module is used to take the one-dimensional initial velocity structure model as a reference, the initial S-wave velocity model as the inversion initial model, integrate the active source high-frequency surface wave dispersion data, the passive source surface wave dispersion data and the single HVSR curve for joint inversion, obtain the three-dimensional fine S-wave velocity structure, and output the final three-dimensional fine velocity structure model in combination with the three-dimensional absolute P-wave velocity model. Specifically, converting the three-dimensional absolute P-wave velocity model into an initial S-wave velocity model includes: For areas within the engineering site covered by the borehole data, the P-wave velocity corresponding to the three-dimensional absolute P-wave velocity model is converted into S-wave velocity using the measured P-wave velocity to S-wave velocity ratio from the borehole data. For unknown lithological areas within the engineering site that are not covered by the borehole data, an empirical conversion formula for transverse and longitudinal wave velocities is used to convert the P-wave velocity corresponding to the three-dimensional absolute P-wave velocity model into S-wave velocity. The acquisition of active source high-frequency surface wave dispersion data specifically includes: High-density linear flow arrays were deployed in the target areas of interest at the engineering site to conduct artificially excited active source surface wave exploration. After correcting the active source earthquake time, the vertical component waveforms were arranged and organized according to the epicentral distance. The frequency-Bessel transform method is used to extract the Rayleigh wave phase velocity dispersion curve from the processed waveform data. The Rayleigh wave phase velocity dispersion curve after quality screening is used as the high-frequency surface wave dispersion data of the active source. Obtaining passive source surface wave dispersion data specifically includes: Continuous background noise observation is conducted based on a dense array of stations deployed in the engineering site area. The continuous noise data collected by each station is preprocessed, including instrument response removal, trend removal, mean removal, normalization, and bandpass filtering. Based on the preprocessed continuous noise data, the noise cross-correlation function between the same components between every two stations is calculated, including vertical-vertical components and tangential-tangential components. The noise cross-correlation functions of multiple time periods are superimposed to extract the empirical Green's function. The frequency-Bessel transform method is used to extract the dispersion curves of Rayleigh waves and Love waves from the empirical Green's function. After quality control and screening of the extracted dispersion curves, the dispersion data of the passive source surface wave is obtained. Obtain the individual HVSR curves for each station in the engineering site area, specifically including: Based on the three-component noise data collected by a single station, the spectral ratio of the horizontal component to the vertical component of a single station is calculated to obtain the HVSR curve of the corresponding station. The peak frequency of the single HVSR curve corresponds to the S-wave resonance frequency of the site strata, and is used to provide constraints on the thickness of the sedimentary layer and the burial depth of the bedrock interface for joint inversion.

Images