A Method for Inverting Lunar Residue Layer Structure and Physical Properties Based on Radar Penetration Detection

By combining radar penetration testing with surface calibration experiments and numerical simulations, the problem of obtaining mechanical parameters in lunar subsurface exploration was solved, achieving high-precision layered structure identification and lateral coverage, supporting lunar base site selection and road assessment.

CN122962678BActive Publication Date: 2026-11-17WUHAN UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202611464068.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-09-20
Publication Date
2026-11-17
Estimated Expiration
2046-09-20

AI Technical Summary

Technical Problem

Existing lunar subsurface exploration technologies cannot effectively obtain key mechanical parameters such as internal friction angle and cohesion. Furthermore, single exploration methods are insufficient to achieve high-precision identification of layered structures and lateral coverage, failing to fully utilize the complementary advantages of radar and penetrating detection.

Method used

A combined radar-penetration method was adopted to construct a lunar soil property database through surface calibration experiments, discrete element simulation, and numerical simulation. By combining ground-penetrating radar and static penetration data, the synergistic inversion of lunar soil layer structure and physical property parameters was achieved, including mechanical property tests, static penetration tests, and ground-penetrating radar scans. Discrete element method simulation and Bayesian method were used for data matching and extrapolation.

Benefits of technology

It enables large-scale, high-precision collaborative acquisition of lunar soil layer structure and physical properties, reduces dependence on heavy equipment and energy, supports non-destructive time-lapse monitoring, and provides key data support for lunar base site selection and lunar surface road bearing capacity assessment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122962678B_ABST
    Figure CN122962678B_ABST
Patent Text Reader

Abstract

This invention discloses a method for inverting lunar regolith layering structure and physical parameters based on a combination of radar penetration testing (GPR) and lunar geological exploration, relating to the fields of deep space exploration and lunar geological surveying. The method constructs a lunar regolith property database by recording the results of surface calibration experiments simulating lunar regolith; it obtains a two-dimensional continuous profile and porosity distribution of the target area, iteratively updating the initial two-dimensional stratigraphic interface and initial porosity distribution; then, using the corrected porosity distribution as a background trend, the CPT cone penetration resistance is extrapolated to the two-dimensional profile of the entire survey line; finally, at each coordinate location, normalized cone penetration resistance and porosity ratio are used as observation pairs, and global matching is performed using the lunar regolith property database to output the physical parameters at that location. This invention effectively integrates the lateral coverage advantage of GPR with the vertical precision advantage of CPT, and can be applied to improve the wide-area detection accuracy and reliability of lunar subsurface structure and mechanical parameters, providing strong support for lunar engineering exploration.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the fields of deep space exploration and lunar geological survey technology, and in particular to a method for inverting the layered structure and physical properties of lunar soil based on a combination of radar penetration and detection. Background Technology

[0002] In-situ exploration of the lunar subsurface structure and physical properties is a core foundation for lunar base site selection, landing stability assessment, rover mobility analysis, and in-situ resource utilization. Current lunar subsurface exploration primarily relies on single geophysical methods such as ground-penetrating radar (GPR), but this has significant limitations: First, there is no direct physical correlation between electromagnetic response and the mechanical behavior of materials, making it impossible to obtain key mechanical parameters such as internal friction angle and cohesion. Second, GPR layering relies on the reflection of electromagnetic waves at the medium interface; the vertical resolution is theoretically limited by the wavelength corresponding to the antenna's center frequency. In practice, high-frequency components attenuate rapidly in the lunar regolith, resulting in the ability to identify only thicker stratigraphic units. It is difficult to effectively distinguish thin layers such as centimeter-level topsoil or gravel interlayers. Furthermore, when the interface is highly undulating, reflection phase axis tracking is prone to cumulative errors, leading to significant uncertainty in layer thickness inversion.

[0003] Static cone penetration testing (CPT) obtains the penetration resistance at the cone tip and the frictional resistance of the sidewalls through continuous penetration, directly estimating soil mechanical parameters and fine-grained stratification information, and has gradually become an important technical means for in-situ mechanical exploration of the lunar surface. However, the detection range of CPT is limited to a single point, resulting in low coverage efficiency and difficulty in reflecting the lateral variation of subsurface parameters. More importantly, existing CPT inversion models for lunar regolith mostly use empirical formulas for terrestrial sandy soils, failing to fully consider the influence of unique lunar regolith characteristics such as sharp-edged particle morphology, extremely low confining pressure, and significant interparticle van der Waals forces caused by ultra-vacuum environments on penetration resistance, thus limiting the applicability of the models.

[0004] Although GPR and CPT are naturally complementary in terms of detection dimensions, with GPR providing continuous lateral coverage and CPT providing high-precision vertical characterization, current studies mostly interpret the results of the two independently or make simple comparisons. They have not yet formed a synergistic inversion framework with a unified physical mechanism as the link, and have failed to fully explore their complementary advantages.

[0005] Therefore, there is an urgent need to develop a joint inversion method based on radar-penetration to achieve large-scale, high-precision collaborative acquisition of lunar subsurface structure and physical parameters. Summary of the Invention

[0006] To construct a joint inversion database based on experimental calibration and numerical simulation expansion, overcome the inherent limitations of single detection methods, and provide reliable technical support for lunar engineering exploration, this invention proposes a method for inverting lunar regolith layering structure and physical property parameters based on a combination of radar penetration and detection. This invention is specifically implemented through the following technologies.

[0007] This invention provides a method for inverting lunar soil layering structure and physical properties based on combined radar penetration and detection, comprising the following steps:

[0008] Surface calibration tests were conducted on simulated lunar soil with different gradations and porosities to obtain corresponding macroscopic mechanical response results, penetration response results, and radar response results; the surface calibration tests included mechanical property tests, static penetration tests, and ground-penetrating radar scanning tests.

[0009] The mechanical property test and static cone penetration test were simulated using the discrete element method to construct a lunar soil property database.

[0010] Initial ground-penetrating radar (GPR) detection and inversion were performed on the target area on the lunar surface to obtain the initial two-dimensional stratigraphic interface and initial porosity distribution. CPT penetration points were selected in the target area on the lunar surface to conduct static cone penetration detection to obtain the one-dimensional fine-layered interface and calculate the corresponding normalized cone penetration resistance.

[0011] The initial porosity distribution is updated using the stratification points on the one-dimensional fine stratification interface as constraints.

[0012] Using the updated porosity distribution as the background trend field, and the normalized cone tip penetration resistance of each layer corresponding to the CPT penetration point in the target area of ​​the lunar surface as hard data, the cone tip penetration resistance is extrapolated along the survey line to construct a two-dimensional normalized cone tip penetration resistance.

[0013] Based on the two-dimensional normalized cone tip penetration resistance and the updated porosity distribution, for each point in the lunar profile formed by the ground-penetrating radar scanning the lunar surface along the survey lines, the lunar soil property database is used for global matching to invert the corresponding physical property parameters, and output the target area on the lunar surface and the two-dimensional continuous distribution of physical property parameters.

[0014] Furthermore, the simulated lunar soils with different gradations and porosities are specifically: simulated lunar soils with different median particle size, fine particle content, ilmenite content, porosity, particle density, and bulk density.

[0015] Furthermore, the method for the mechanical property test is as follows: triaxial compression tests are conducted on the simulated lunar soil with different gradations and void ratios, and the gradation characteristic parameters of the simulated lunar soil and the macroscopic mechanical response results under different void ratios are recorded; the gradation characteristic parameters include median particle size and fine particle content, and the macroscopic mechanical response results include critical void ratio, internal friction angle and cohesion, and stress-strain curves are obtained.

[0016] Furthermore, the method of the static cone penetration test is as follows: the simulated lunar soil with different gradations and porosities is subjected to cone tip penetration at a constant rate, and the cone tip penetration resistance-penetration depth variation curve is recorded; the cone tip penetration resistance is normalized to obtain the normalized cone tip penetration resistance; with the state parameters as independent variables and the normalized cone tip penetration resistance as dependent variables, the relationship between the gradation characteristic parameters, porosity and normalized cone tip penetration resistance of the simulated lunar soil with different gradations and porosities is fitted to establish the relationship.

[0017] Furthermore, the ground-penetrating radar scanning experiment involves: conducting co-offset data acquisition on the simulated lunar soil with different gradations and porosities, and inverting the dielectric constant and loss tangent; using the relationship between ilmenite content and particle density, the relationship between dielectric constant and bulk density, the relationship between ilmenite content and loss tangent, and the calculation formula for porosity, to obtain the relationship between dielectric constant, loss tangent, and porosity of the simulated lunar soil.

[0018] Furthermore, the method for constructing a lunar soil properties database by simulating the aforementioned mechanical property tests and static cone penetration tests using the discrete element method is as follows:

[0019] The mechanical property test and static cone penetration test were simulated using the discrete element method to obtain the corresponding numerical simulation results.

[0020] By comparing the numerical simulation results with the macroscopic mechanical response results and the penetration response results, the particle contact model parameters in the discrete element numerical simulation are calibrated.

[0021] Numerical simulation samples with different gradations and porosity combinations were generated in batches, and the macroscopic mechanical simulation response and penetration response simulation results were obtained under the lunar environment.

[0022] The macroscopic mechanical simulation response results and the penetration response simulation results are stored in the lunar soil property database; each numerical simulation sample in the lunar soil property database includes gradation characteristic parameters, porosity, normalized cone tip penetration resistance, critical porosity, internal friction angle and cohesion.

[0023] Furthermore, the method for obtaining the initial two-dimensional stratigraphic interface and initial porosity distribution by performing initial ground-penetrating radar detection and initial inversion on the target area on the lunar surface is as follows:

[0024] Ground-penetrating radar is used to collect radar data from target areas on the lunar surface using co-offset sampling, and the radar data obtained from the scan is preprocessed.

[0025] Based on the preprocessed radar data, an initial two-dimensional stratigraphic interface is established using the full waveform inversion method to obtain the dielectric constant and loss tangent; the dielectric constant is then converted into the initial porosity distribution of the target region on the lunar surface.

[0026] Furthermore, the preprocessing includes: zero-time correction, background removal, and gain compensation.

[0027] Furthermore, by selecting CPT penetration points in the target area on the lunar surface and conducting static cone penetration tests, a one-dimensional fine-layered interface was obtained. The method for calculating the corresponding normalized cone tip penetration resistance is as follows:

[0028] Based on radar scanning of the lunar surface along survey lines, the initial lunar profile areas were obtained by screening, and CPT penetration points were selected.

[0029] At the CPT penetration point, the CPT probe was inserted at a constant rate, and the cone tip penetration resistance-penetration depth variation curve was recorded.

[0030] Using the Bayesian method, statistical analysis was performed on the cone tip penetration resistance-penetration depth variation curve to identify several one-dimensional fine-layer interfaces and calculate the corresponding normalized cone tip penetration resistance.

[0031] Furthermore, the method for updating the initial porosity distribution, using the stratification points on the one-dimensional fine stratification interface as constraints, is as follows:

[0032] For each CPT penetration point, the initial two-dimensional stratigraphic interface and the one-dimensional fine stratification interface are compared to construct an objective function for interface depth difference.

[0033] The initial porosity distribution is adjusted using the interface depth difference objective function, and the above process is iterated until the interface depth difference objective function equals 0, resulting in the updated porosity distribution.

[0034] Furthermore, using the updated porosity distribution as the background trend field, and the normalized cone tip penetration resistance of each layer corresponding to the CPT penetration point in the target area of ​​the lunar surface as hard data, the cone tip penetration resistance is extrapolated along the survey line to construct the two-dimensional normalized cone tip penetration resistance.

[0035] At each of the CPT penetration points, based on the updated GPR porosity and its lateral gradient, an external drift matrix is ​​constructed, and the drift vector of the point to be estimated is calculated.

[0036] Define the experimental variation function;

[0037] By fitting the experimental variation function with a spherical model, the nugget value, sill value and experimental variation function are obtained, and the covariance matrix between CPT penetration points and the covariance vector between the point to be estimated and the CPT penetration point are constructed.

[0038] Construct and solve the Kriging equations to obtain the optimal weight of each CPT penetration point to any point to be estimated.

[0039] The optimal weight is weighted and summed with the measured normalized cone tip penetration resistance at each of the CPT penetration points to obtain the optimal extrapolated value of the cone tip penetration resistance at the point to be estimated, thus forming a two-dimensional normalized cone tip penetration resistance.

[0040] Furthermore, based on the two-dimensional normalized cone tip penetration resistance and the updated porosity distribution, for each point in the lunar profile formed by ground-penetrating radar scanning the lunar surface along survey lines, the lunar soil property database is used for global matching to invert the corresponding physical property parameters, and the method for outputting the target area and the two-dimensional continuous distribution of physical property parameters on the lunar surface is as follows:

[0041] Based on the updated porosity distribution and the two-dimensional normalized cone penetration resistance, at each point in the lunar profile formed by the ground-penetrating radar scanning the lunar surface along the survey lines, the cone penetration resistance and porosity are used as observation pairs to perform global matching in the lunar soil property database to complete the optimal matching.

[0042] Calculate the relative error between the predicted cone tip penetration resistance and the measured cone tip penetration resistance for each sample, as well as the relative error between the predicted porosity and the updated porosity distribution.

[0043] The sample with the smallest comprehensive error is taken as the inversion result. The corresponding gradation characteristic parameters, critical porosity, internal friction angle and cohesion are inverted. Finally, the two-dimensional continuous distribution of the layered structure and physical property parameters of the target area along the full profile of the survey line on the actual lunar surface is output.

[0044] Compared with the prior art, the advantages of the present invention are:

[0045] 1. This method records the macroscopic mechanical response and penetration response of lunar soil under different lunar soil gradation and porosity characteristics by combining the lateral continuous coverage of ground penetrating radar (GPR) and the vertical high-precision mechanical test of static cone penetration (CPT) with discrete element numerical simulation method. It also constructs a lunar soil property database that considers the special environment of the lunar surface, and realizes the accurate mapping of discrete mechanical parameters to wide-area continuous physical property distribution.

[0046] 2. The dry lunar environment results in extremely low dielectric loss, allowing ground-penetrating radar to obtain high-fidelity continuous profiles tens of meters or even deeper. Only a very small amount of static penetration testing is needed for mechanical calibration, which can convert wide-area electromagnetic information into mechanical parameter distributions such as cone tip penetration resistance. This significantly reduces the dependence on heavy penetration equipment, reaction devices, and energy supply. Under the limited carrying capacity of the lander, it breaks through the limitation of sparse sampling points and achieves large-scale, high-precision mechanical characterization of the shallow lunar surface.

[0047] 3. The method provided by this invention supports time-shift monitoring based on non-destructive radar scanning, which can track the evolution of lunar soil compaction and changes in foundation stability, providing key data support for lunar base site selection, lunar surface road bearing capacity assessment, and long-term safe operation. Attached Figure Description

[0048] Figure 1 Here is a flowchart illustrating the method for inverting the layered structure and physical properties of lunar soil provided in the embodiment.

[0049] Figure 2 For example, in the lunar soil layering structure and physical property parameter inversion method provided in the embodiment, step two is the macroscopic mechanical simulation response result and the penetration response simulation result obtained after mechanical property test and static penetration test.

[0050] Figure 3 For example, in the lunar soil layering structure and physical property parameter inversion method provided in the embodiment, the updated porosity distribution e obtained in step five is... GPR (x,z).

[0051] Figure 4 For example, in the lunar soil layering structure and physical property parameter inversion method provided in the embodiment, the normalized cone tip penetration resistance Q of each layer at the CPT penetration point in step four is... t A schematic diagram of the change curve of penetration depth.

[0052] Figure 5 For example, in the lunar soil layering structure and physical property parameter inversion method provided in the embodiment, the two-dimensional normalized cone tip penetration resistance Q constructed after extrapolation in step six is... t (x,z) distribution diagram.

[0053] Figure 6 This is a schematic diagram of the process of global matching in the lunar soil property database using two-dimensional normalized cone tip penetration resistance and updated porosity distribution. Detailed Implementation

[0054] The technical solution of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] Example

[0056] The method for inverting lunar soil layering structure and physical property parameters based on radar penetration combined with other methods provided in this embodiment is as follows: Figure 1 As shown, the steps include the following.

[0057] Step 1: Using simulated lunar soil with different gradations and porosities, conduct surface calibration tests (mechanical property tests, static cone penetration tests, and radar scanning tests) to obtain the macroscopic mechanical response, cone penetration response, and radar response results of simulated lunar soil with different gradations and porosities.

[0058] 1. Preparation of simulated lunar soil

[0059] Basalt and ilmenite, which have mineral compositions similar to those of lunar soil from Chang'e 5, were selected as raw materials. The sharp morphology of lunar soil particles was simulated by controlling the crushing and rounding process.

[0060] Sieving and compounding to prepare particles with different median particle sizes d 50 The simulated lunar soil samples contained fine particle content (S, i.e., the mass percentage of particles with a diameter less than 48 μm) and ilmenite content (F). Specific parameter information for the simulated lunar soil samples is shown in Table 1 below.

[0061] Table 1: Specific parameters of simulated lunar soil samples with different gradations and void ratios

[0062] 2. Mechanical property test

[0063] For each gradation-void ratio of the simulated lunar soil samples in Table 1, triaxial compression tests were conducted to determine their mechanical properties. The gradation characteristic parameters (median particle size d) of each simulated lunar soil sample were recorded. 50 The results of the fine particle content (S) and the macroscopic mechanical response under different porosity ratios (including the critical porosity e) are also presented. c (Internal friction angle and cohesion) to obtain stress-strain curves.

[0064] 3. Static cone penetration test (CPT)

[0065] (1) Static cone penetration tests were conducted on simulated lunar soil samples of each gradation-porosity ratio in Table 1.

[0066] Cone tip penetration was performed at a constant rate in simulated lunar soil samples with known porosity, and the cone tip penetration resistance-penetration depth variation curves were recorded.

[0067] (2) The cone tip penetration resistance is normalized using Equation I to obtain the normalized cone tip penetration resistance Q. t ;

[0068] Formula I;

[0069] In the formula, q t For the cone tip penetration resistance, σ v0 σ' is the total overburden pressure of the soil. v0 This refers to the effective overburden pressure of the soil.

[0070] (3) Using the state parameter ψ as the independent variable and the normalized cone tip penetration resistance Q as the independent variable. t Using the dependent variable, a normalized empirical relationship between cone tip penetration resistance and state parameters was established through fitting, thereby establishing the gradation characteristic parameters (median particle size d) of different simulated lunar soil samples. 50 Fine particle content (S), void ratio, and normalized cone tip penetration resistance (Q) t The relationship is shown in equations II to V.

[0071] Formula II;

[0072] Formula III;

[0073] Formula IV;

[0074] Formula V;

[0075] In the formula, K and m are fitting parameters, e is the porosity, and e c The critical void ratio (obtained through triaxial compression test).

[0076] 4. Ground Penetrating Radar (GPR) Scanning Test

[0077] (1) Ground penetrating radar was used under various known conditions (porosity e, ilmenite content F, particle density ρ shown in Table 1). s Bulk density ρ v The simulated lunar regolith surface was sampled using a common offset method.

[0078] (2) Based on radar B-scan data, the dielectric constant α and loss tangent tanδ are inverted.

[0079] (3) Titanium iron content F-particle density ρ based on formula VI s The relationship between the dielectric constant α and the volume density ρ in equation VII. v The relationship between the ilmenite content F and the loss tangent tanδ in Equation VIII, and the porosity e in Equation IX are obtained. Finally, the relationship between the dielectric constant ɛ and the loss tangent tanδ and the porosity of the simulated lunar soil is obtained.

[0080] Formula VI;

[0081] Formula VII;

[0082] Formula VIII; Formula IX.

[0083] Step 2: Simulate the surface calibration experiment using the discrete element method to construct a lunar soil properties database.

[0084] 1. Conduct discrete element numerical simulation: Use the discrete element method to simulate the mechanical property test (triaxial compression test) and static cone penetration test (CPT) process in step one, and obtain the corresponding numerical simulation results.

[0085] The discrete element method used in this embodiment is specifically a three-dimensional discrete element method for numerical simulation. Considering the significant van der Waals forces between lunar soil particles in a vacuum environment, a linear rolling resistance contact model with adhesion is used as the inter-particle contact model.

[0086] In the triaxial compression test simulation, a cubic specimen with 5000 particles was generated. A constant confining pressure (6, 10, 15 kPa) was maintained by a wall servo mechanism, and an axial load was applied at a constant strain rate of 0.5% / min. The stress-strain curves were recorded.

[0087] In the static cone penetration test (CPT) simulation, a rigid wall assembly is used as the model box (considering the size effect, the side length of the model box is not less than 800×d). 50 Modeling with a cone-shaped probe (cone angle 60°, base area 10 cm²) 2 The model box was filled with lunar soil particles and penetrated at a constant rate of 10 mm / s. The cone tip penetration resistance-penetration depth variation curve was recorded.

[0088] 2. By comparing the above numerical simulation results with the macroscopic mechanical response results (i.e., stress-strain curves) and penetration response results (i.e. cone tip penetration resistance-penetration depth variation curves) obtained in step one, the particle contact model parameters (elastic modulus, stiffness ratio, friction coefficient) in the discrete element numerical simulation are calibrated.

[0089] The specific parameters of the particle contact model calibrated in this embodiment are: elastic modulus 4 GPa, stiffness ratio 2, and friction coefficient 0.76.

[0090] 3. Through orthogonal design, numerical simulation samples with different gradations and porosity combinations are generated in batches to simulate mechanical properties and static penetration tests under special lunar environments (such as low gravity, ultra-vacuum, etc., which can be achieved by setting the corresponding lunar gravity and interparticle van der Waals forces), and obtain the corresponding macroscopic mechanical simulation response results and penetration response simulation results.

[0091] The process of generating numerical simulation samples using orthogonal design is as follows: In orthogonal design, the median particle size d 50 Three factors were considered: fine particle content (S) and porosity (e). Each factor was assigned three levels to construct an L9(3) model. 3 An orthogonal array was used to generate a total of 9 sets of numerical simulation samples. Among them, d 50The three levels of particle size distribution were set to 0.05 mm, 0.1 mm, and 0.15 mm, covering the fine, medium, and coarse-grained lunar soil ranges, respectively; the three levels of fine particle content (S) were 10%, 35%, and 60%, respectively; and the three levels of void ratio (e) were set to 0.45, 0.7, and 1.1, corresponding to dense, medium-dense, and loose states, respectively. Based on the target gradation and pore characteristics determined by each orthogonal combination, corresponding particle aggregates were generated, and the final void ratio was controlled to reach the target value (allowable deviation ±0.05) through stratified under-compaction.

[0092] The corresponding macroscopic mechanical simulation response results and penetration response simulation results are as follows: Figure 2 As shown.

[0093] 4. Store the macroscopic mechanical simulation response results and the penetration response simulation results into the lunar soil property database D shown in Equation X.

[0094] Each numerical simulation sample in the lunar soil properties database includes: gradation characteristic parameters (d 50 S), void ratio e, normalized cone tip penetration resistance Q t Critical porosity e c The internal friction angle φ and the cohesive force c.

[0095] Formula X.

[0096] Step 3: Perform initial ground-penetrating radar (GPR) detection and initial inversion on the target area on the lunar surface to obtain the initial two-dimensional stratigraphic interface and initial porosity distribution.

[0097] 1. Preprocess the radar data obtained from the scan.

[0098] Ground-penetrating radar (GPR) was used to acquire radar data from a target area on the actual lunar surface using co-offset methods. The acquired radar data was then preprocessed, including zero-time correction, background removal, and gain compensation.

[0099] In this embodiment, the zero-time correction adopts the direct wave method to determine the accurate start time of the radar signal and eliminate the signal propagation delay caused by system hardware (such as cables and circuits).

[0100] In this embodiment, background removal employs a mean subtraction method to suppress or eliminate background clutter and noise (including direct coupling waves between antennas, strong surface reflections, and environmental electromagnetic interference) in radar data, thereby improving the signal-to-noise ratio and enhancing the identifiability of weak effective reflection signals from underground targets.

[0101] In this embodiment, the gain compensation adopts the diffusion and exponential compensation (SEC) method to compensate for the energy attenuation caused by spherical diffusion and medium absorption during the propagation of electromagnetic waves in the underground medium, restore the reflected signal intensity of deep targets, and enable both deep and shallow targets to be clearly displayed on the radar profile, thereby improving the overall imaging quality.

[0102] 2. Based on the preprocessed radar data described above, the initial two-dimensional stratigraphic interface H is established using the full waveform inversion method. GPR0 (x,z) is obtained, and the dielectric constant ɛ and the loss tangent tanδ are obtained.

[0103] The full waveform inversion method is as follows: Based on preprocessed scanning ground-penetrating radar data, a two-dimensional full waveform inversion is employed. By iteratively minimizing the waveform residuals between the scanning ground-penetrating radar data and the numerically forward-modeled radar data, the spatial distribution of the dielectric constant α and the loss tangent tanδ is inverted. The numerically forward-modeling uses a Maxwell's equations solver based on the finite-difference time-domain method.

[0104] In (x,z), x represents the abscissa of the lunar surface scanned by the ground penetrating radar along the survey line (i.e., the predetermined path of the radar scan), and z represents the depth at which the ground penetrating radar scans the lunar surface along the survey line and forms a lunar profile.

[0105] 3. Convert the dielectric constant α into the initial porosity distribution e of the target region on the lunar surface using equations VI to IX from step one. 0GPR (x,z).

[0106] Step 4: Select CPT penetration points in the target area on the lunar surface, conduct static cone penetration (CPT) exploration to obtain a one-dimensional fine layered interface, and calculate the corresponding normalized cone tip penetration resistance.

[0107] 1. Based on the initial lunar profile formed by ground-penetrating radar scanning the lunar surface along the survey lines, areas with uniform reflection characteristics and no signs of diffraction from large rocks are selected as CPT penetration points. i .

[0108] 2. Insert the CPT probe at a constant rate and record the curve of cone tip penetration resistance versus penetration depth.

[0109] 3. Using Bayesian methods, statistical analysis was performed on the cone tip penetration resistance-penetration depth variation curves to identify several one-dimensional fine-grained stratification interfaces (i.e., dividing the initial lunar profile into several layers based on the stratigraphic depth obtained from the initial lunar profile scan). CPT (z k ), k=1,2,…,K, and use Equation I to calculate the normalized cone tip penetration resistance Qt, as follows Figure 3 As shown.

[0110] Step 5: Update the initial porosity distribution using the layering points on the one-dimensional fine layering interface as constraints.

[0111] The one-dimensional fine-layered interface H obtained in step four CPT (z k Using the stratified points on the surface as constraints, update the initial porosity distribution e obtained in step three. 0GPR (x,z), making the initial two-dimensional stratigraphic interface H in GPR inversion GPR0 Layered points on (x,z) and one-dimensional fine-layered interface H CPT (z k The CPT penetration points are consistent, and the updated porosity distribution e is obtained. GPR (x,z).

[0112] 1. For each CPT penetration point s i The initial two-dimensional stratigraphic interface H of the ground-penetrating radar GPR obtained in step three is used to define the layered interface. GPR0 (x,z), and the one-dimensional fine-layered interface H of the static penetration test (CPT) obtained in step four. CPT (z k By comparing the two, we construct the interface depth difference objective function shown in Equation IX, as shown in Equation XⅠ.

[0113] Formula XⅠ;

[0114] Where Φ is the objective function for interface depth difference, z GPR Let k be the depth of the GPR layered interface, and z be the k-th layer. CPT Determine the layered interface depth for CPT.

[0115] 2. Adjust the initial void ratio distribution e using Φ as the objective function. 0GPR (x,z), iterate the above process until Φ equals 0. That is, make the layered interface obtained by GPR inversion consistent with the layered interface of CPT at the constraint position, and obtain the updated porosity distribution e. GPR (x,z), such as Figure 2 As shown.

[0116] Step 6: Using the updated porosity distribution e GPR Using (x,z) as the background trend field, and taking the normalized cone tip penetration resistance Qt corresponding to each layer of the CPT penetration point in the target area on the lunar surface as hard data, the cone tip penetration resistance is extrapolated along the survey line to construct a two-dimensional normalized cone tip penetration resistance Qt. t (x,z).

[0117] The updated porosity distribution e obtained in step five GPR (x,z) (e.g.) Figure 3As shown in the figure, the background trend field is used, and the normalized cone tip penetration resistance Q of each layer at the CPT penetration point is used. t (like Figure 4 As shown in the figure, using the external drift kriging method as hard data, the normalized cone tip penetration resistance is extrapolated along the survey line to construct a two-dimensional normalized cone tip penetration resistance Q. t (x,z), such as Figure 5 As shown.

[0118] 1. At each CPT penetration point s i At this point, based on the updated GPR porosity e GPR (s i ) and its lateral gradient Construct the external drift matrix F and calculate the drift vector f0 at the point to be estimated s0 (i.e., the position on the two-dimensional profile where there is no CPT result).

[0119] To quantify the spatial autocorrelation characteristics of cone tip penetration resistance within the lunar regolith, the experimental variogram γ(h) is defined using the following equation XII:

[0120] Formula XII;

[0121] Where h is the spatial lag distance, and N(h) is the initial two-dimensional stratigraphic interface H. GPR0 The total number of all possible pairs of points that are h apart on (x,z).

[0122] 2. By fitting γ(h) with a spherical model, the nugget value, sill value and experimental variation parameters are obtained, and then the covariance matrix C between CPT penetration points and the covariance vector c0 between the point to be estimated and the CPT penetration point are constructed.

[0123] 3. To minimize the estimation variance under unbiased conditions, the Kriging equations shown in Equations VIII and XIV are constructed and solved to obtain the optimal weight ω of each CPT penetration point to any estimated point s0. i .

[0124] Formula XIII;

[0125] w=[ω1,ω2,...,ω N ] T Formula XⅣ;

[0126] Where T is the matrix transpose symbol, w is the weight coefficient vector of each CPT point to be determined, and μ is the Lagrange multiplier vector used to force the unbiased condition to be met.

[0127] 4. As shown in equation XV below, the optimal weight ω i The measured normalized cone tip penetration resistance Q at each CPT penetration point t(s i By performing a weighted summation, the optimal extrapolated value of the cone tip penetration resistance at the point s0 to be estimated is obtained. The optimal extrapolated value at each point constitutes the two-dimensional normalized cone tip penetration resistance Q. t (x,z).

[0128] Formula XV.

[0129] Step 7: Based on the two-dimensional normalized cone tip penetration resistance and the updated porosity distribution, for each point in the lunar profile formed by the ground-penetrating radar scanning the lunar surface along the survey line, global matching is performed using the lunar soil property database to invert the corresponding physical property parameters, and the target area on the lunar surface and the two-dimensional continuous distribution of physical property parameters are output.

[0130] 1. Based on the updated porosity distribution e obtained in step five GPR (x,z), and the two-dimensional normalized cone tip penetration resistance Q obtained in step six. t (x,z), at each point (x,z) in the lunar profile formed by ground-penetrating radar scanning the lunar surface along the survey lines, with (Q t e) represents the observation pair, which undergoes global matching within the lunar soil property database D constructed in step two to achieve optimal matching, such as... Figure 6 As shown;

[0131] 2. Calculate the predicted cone tip penetration resistance q for each observation pair. c (Right now Figure 5 Step III of the inversion approach (calculated based on the lunar soil property database) and the measured Q t (At the specific CPT penetration point, the measured Q) t The relative error M (where other positions are the optimal extrapolated values) q And the predicted porosity ratio and e GPR (Updated porosity distribution e) GPR The relative error M of (x,z) e .

[0132] 3. Take the comprehensive error (M) q +M e The smallest sample is used as the inversion result, and the corresponding gradation characteristic parameters (median particle size d) are inverted. 50 Fine particle content (S) and critical void ratio (e) c The internal friction angle φ and cohesion c are used to ultimately output the two-dimensional continuous distribution of the layered structure and physical properties of the target area on the actual lunar surface along the survey line (i.e., the initial lunar profile formed by the ground penetrating radar scanning the lunar surface along the survey line in step three).

[0133] The above detailed embodiments describe the implementation of the present invention; however, the present invention is not limited to the specific details described in the above embodiments. Within the scope of the claims and technical concept of the present invention, various simple modifications and changes can be made to the technical solution of the present invention, and these simple modifications all fall within the protection scope of the present invention.

Claims

1. A method for inverting lunar soil layering structure and physical property parameters based on radar penetration detection, characterized in that: step: Surface calibration tests were conducted on simulated lunar soil with different gradations and porosities to obtain corresponding macroscopic mechanical response results, penetration response results, and radar response results; the surface calibration tests included mechanical property tests, static penetration tests, and ground-penetrating radar scanning tests. The mechanical property test and static cone penetration test were simulated using the discrete element method to construct a lunar soil property database. Initial ground-penetrating radar detection and initial inversion were performed on the target area on the lunar surface to obtain the initial two-dimensional stratigraphic interface and initial porosity distribution. Select CPT penetration points in the target area on the lunar surface, conduct static cone penetration tests, obtain one-dimensional fine-layered interfaces, and calculate the corresponding normalized cone tip penetration resistance. The initial porosity distribution is updated using the stratification points on the one-dimensional fine stratification interface as constraints. Using the updated porosity distribution as the background trend field, and the normalized cone tip penetration resistance of each layer corresponding to the CPT penetration point in the target area of ​​the lunar surface as hard data, the cone tip penetration resistance is extrapolated along the survey line to construct a two-dimensional normalized cone tip penetration resistance. Based on the two-dimensional normalized cone tip penetration resistance and the updated porosity distribution, for each point in the lunar profile formed by the ground-penetrating radar scanning the lunar surface along the survey lines, the lunar soil property database is used for global matching to invert the corresponding physical property parameters, and output the target area on the lunar surface and the two-dimensional continuous distribution of physical property parameters.

2. The method for inverting lunar soil layering structure and physical property parameters based on radar penetration testing as described in claim 1, characterized in that, The simulated lunar soils with different gradations and void ratios are specifically: simulated lunar soils with different median particle size, fine particle content, ilmenite content, void ratio, particle density, and bulk density.

3. The method for inverting lunar soil layering structure and physical property parameters based on radar penetration testing as described in claim 1, characterized in that, The method for the mechanical property test is as follows: triaxial compression tests are conducted on the simulated lunar soil with different gradations and void ratios, and the gradation characteristic parameters and macroscopic mechanical response results under different void ratios of the simulated lunar soil are recorded; the gradation characteristic parameters include median particle size and fine particle content, and the macroscopic mechanical response results include critical void ratio, internal friction angle and cohesion, and stress-strain curves are obtained; The method of the static cone penetration test is as follows: the simulated lunar soil with different gradations and porosities is penetrated at a constant rate with a cone tip, and the cone tip penetration resistance-penetration depth variation curve is recorded; the cone tip penetration resistance is normalized to obtain the normalized cone tip penetration resistance. Using state parameters as independent variables and normalized cone tip penetration resistance as dependent variables, we fitted and established the relationship between the gradation characteristic parameters, porosity and normalized cone tip penetration resistance of the simulated lunar soil with different gradations and porosities. The ground-penetrating radar scanning experiment involves: collecting data at common offset on simulated lunar soil with different gradations and porosities, and inverting the dielectric constant and loss tangent; using the relationship between ilmenite content and particle density, the relationship between dielectric constant and bulk density, the relationship between ilmenite content and loss tangent, and the calculation formula for porosity, obtaining the relationship between dielectric constant, loss tangent, and porosity of simulated lunar soil.

4. The method for inverting lunar soil layering structure and physical property parameters based on radar penetration testing as described in claim 1, characterized in that, The method for constructing a lunar soil properties database by simulating the mechanical property test and static cone penetration test using the discrete element method is as follows: The mechanical property test and static cone penetration test were simulated using the discrete element method to obtain the corresponding numerical simulation results. By comparing the numerical simulation results with the macroscopic mechanical response results and the penetration response results, the particle contact model parameters in the discrete element numerical simulation are calibrated. Numerical simulation samples with different gradations and porosity combinations were generated in batches, and the macroscopic mechanical simulation response and penetration response simulation results were obtained under the lunar environment. The macroscopic mechanical simulation response results and the penetration response simulation results are stored in the lunar soil property database; each numerical simulation sample in the lunar soil property database includes gradation characteristic parameters, porosity, normalized cone tip penetration resistance, critical porosity, internal friction angle and cohesion.

5. The method for inverting lunar soil layering structure and physical properties based on radar penetration testing as described in claim 1, characterized in that, The method for obtaining the initial two-dimensional stratigraphic interface and initial porosity distribution by performing initial ground-penetrating radar detection and initial inversion on a target area on the lunar surface is as follows: Ground-penetrating radar is used to collect radar data from target areas on the lunar surface using co-offset sampling, and the radar data obtained from the scan is preprocessed. Based on the preprocessed radar data, the initial two-dimensional stratigraphic interface was established using the full waveform inversion method, and the dielectric constant and loss tangent were obtained. The dielectric constant is converted into the initial porosity distribution of the target region on the lunar surface.

6. The method for inverting lunar soil layering structure and physical properties based on radar penetration testing as described in claim 1, characterized in that, The CPT penetration point in the target area on the lunar surface is selected for static cone penetration testing to obtain a one-dimensional fine-layered interface. The method for calculating the corresponding normalized cone tip penetration resistance is as follows: Based on radar scanning of the lunar surface along survey lines, the initial lunar profile areas were obtained by screening, and CPT penetration points were selected. At the CPT penetration point, the CPT probe was inserted at a constant rate, and the cone tip penetration resistance-penetration depth variation curve was recorded. Using the Bayesian method, statistical analysis was performed on the cone tip penetration resistance-penetration depth variation curve to identify several one-dimensional fine-layer interfaces and calculate the corresponding normalized cone tip penetration resistance.

7. The method for inverting lunar soil layering structure and physical property parameters based on radar penetration testing as described in claim 1, characterized in that, The method for updating the initial porosity distribution, using the stratification points on the one-dimensional fine stratification interface as constraints, is as follows: For each CPT penetration point, the initial two-dimensional stratigraphic interface and the one-dimensional fine stratification interface are compared to construct an objective function for interface depth difference. The initial porosity distribution is adjusted using the interface depth difference objective function, and the above process is iterated until the interface depth difference objective function equals 0, resulting in the updated porosity distribution.

8. The method for inverting lunar soil layering structure and physical property parameters based on radar penetration testing as described in claim 1, characterized in that, Using the updated porosity distribution as the background trend field, and the normalized cone tip penetration resistance of each layer corresponding to the CPT penetration point in the target area of ​​the lunar surface as hard data, the cone tip penetration resistance is extrapolated along the survey line to construct the two-dimensional normalized cone tip penetration resistance. At each of the CPT penetration points, based on the updated GPR porosity and its lateral gradient, an external drift matrix is ​​constructed, and the drift vector of the point to be estimated is calculated. Define the experimental variation function; By fitting the experimental variation function with a spherical model, the nugget value, sill value and experimental variation function are obtained, and the covariance matrix between CPT penetration points and the covariance vector between the point to be estimated and the CPT penetration point are constructed. Construct and solve the Kriging equations to obtain the optimal weight of each CPT penetration point to any point to be estimated. The optimal weight is weighted and summed with the measured normalized cone tip penetration resistance at each of the CPT penetration points to obtain the optimal extrapolated value of the cone tip penetration resistance at the point to be estimated, thus forming a two-dimensional normalized cone tip penetration resistance.

9. The method for inverting lunar soil layering structure and physical properties based on radar penetration testing as described in claim 1, characterized in that, Based on the two-dimensional normalized cone tip penetration resistance and the updated porosity distribution, for each point in the lunar profile formed by ground-penetrating radar scanning the lunar surface along survey lines, the lunar soil property database is used for global matching to invert the corresponding physical property parameters, and the method for outputting the target area and the two-dimensional continuous distribution of physical property parameters on the lunar surface is as follows: Based on the updated porosity distribution and the two-dimensional normalized cone penetration resistance, at each point in the lunar profile formed by the ground-penetrating radar scanning the lunar surface along the survey lines, the cone penetration resistance and porosity are used as observation pairs to perform global matching in the lunar soil property database to complete the optimal matching. Calculate the relative error between the predicted cone tip penetration resistance and the measured cone tip penetration resistance for each sample, as well as the relative error between the predicted porosity and the updated porosity distribution. The sample with the smallest comprehensive error is taken as the inversion result. The corresponding gradation characteristic parameters, critical porosity, internal friction angle and cohesion are inverted. Finally, the two-dimensional continuous distribution of the layered structure and physical property parameters of the target area along the full profile of the survey line on the actual lunar surface is output.

Citation Information

Patent Citations

  • Method for realizing joint survey of sand blowing and filling amount by adopting ground penetrating radar and static penetrometer

    CN103741658A

  • Soil and topography surveying

    US20030083819A1