A Multi-Source Information Fusion Method for Sand Liquefaction Based on Evidence Theory

By integrating SPT-N and Vs information and combining them with random field theory, the limitations of single data sources and spatial variability in existing technologies for sand liquefaction identification are solved, achieving more accurate liquefaction risk assessment and uncertainty quantification, and making it suitable for complex geological conditions.

CN123120080APending Publication Date: 2026-11-17STATE GRID JIANGSU ECONOMIC RES INST +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611260741.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-19
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies for identifying sand liquefaction suffer from limitations such as relying on a single data source, insufficient determinism, and neglect of spatial variability, resulting in insufficient accuracy and an inability to effectively handle measurement uncertainties and spatial variability.

Method used

The Dempster-Shafer evidence theory is used to integrate two types of information: standard penetration test (SPT-N) and shear wave velocity (Vs). The spatial variability of soil parameters is handled by combining random field theory. The probability interval is constructed by confidence degree and likelihood degree to conduct liquefaction risk assessment.

Benefits of technology

It improves the accuracy of discrimination, provides probabilistic output, quantifies uncertainty, and is suitable for liquefaction risk assessment under complex geological conditions, especially for liquefaction risk assessment under complex geological conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123120080A_ABST
    Figure CN123120080A_ABST
Patent Text Reader

Abstract

This invention relates to the interdisciplinary field of geotechnical engineering and earthquake engineering, specifically a multi-source information fusion method for sand liquefaction discrimination based on evidence theory, particularly suitable for liquefaction risk assessment under complex geological conditions. The method includes S1: acquiring a geotechnical historical exploration dataset of the target area, which includes dynamic parameters obtained through in-situ testing techniques and static and physical state parameters obtained through laboratory geotechnical tests; S2: based on SPT-N data and V... s The data is used to calculate the liquefaction safety factor; step S3: information fusion is performed on the liquefaction safety factor using Dempster-Shafer evidence theory; S4: the liquefaction probability interval is calculated; S5: a liquefaction decision is made based on the relationship between the liquefaction confidence level and the non-liquefaction confidence level; this invention innovatively introduces evidence theory to realize the relationship between SPT-N and shear wave velocity V s The fusion of dual evidence.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention generally relates to the fields of geotechnical engineering and earthquake engineering technology. Specifically, it relates to a multi-source information fusion method for sand liquefaction discrimination based on the Dempster-Shafer evidence theory, which is particularly suitable for liquefaction risk assessment under complex geological conditions. Background Technology

[0002] Soil liquefaction is a significant geological hazard in earthquake engineering. It refers to the phenomenon where saturated sand, under seismic loading, experiences a rapid increase in pore water pressure and a decrease in effective stress to zero, leading to the soil losing shear strength and exhibiting liquefied properties. Liquefaction can cause severe consequences such as ground subsidence, lateral expansion, and building tilting and collapse, posing a major threat to engineering safety. Traditional liquefaction assessment methods primarily rely on a single data source (Standard Penetration Test-N or shear wave velocity V). s The following technical defects exist:

[0003] (1) Limitations of a single data source: Existing methods cannot effectively handle measurement uncertainty and spatial variability;

[0004] (2) Insufficient deterministic criteria: The traditional safety factor method provides a binary decision and lacks probabilistic description;

[0005] (3) Spatial variability is ignored: Most methods do not fully consider the relevant distances and non-stationary characteristics of soil parameters.

[0006] In existing technologies, although random field simulation methods can describe the variability of parameter space, they have failed to solve the problem of multi-source information fusion. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method that can effectively integrate SPT-N and V s Multi-source information improves discrimination accuracy, provides probabilistic output, and supports quantified decision-making for uncertainty, making it suitable for liquefaction discrimination methods under complex geological conditions.

[0008] To achieve the above objectives, the technical solution adopted by this invention is: a multi-source information fusion method for sand liquefaction discrimination based on Dempster-Shafer evidence theory, which includes the following steps:

[0009] S1: Obtain the geotechnical history exploration dataset of the target area. This dataset includes dynamic parameters obtained through in-situ testing techniques, as well as static and physical state parameters obtained through laboratory geotechnical tests. The dynamic parameters include Standard Penetration Test (SPT-N) data and shear wave velocity data V. s ;

[0010] S2: Calculate the corresponding liquefaction safety factor F based on the standard penetration test data SPT-N and shear wave velocity data Vs, respectively. SPT and F Vs ;

[0011] S3: The Dempster-Shafer evidence theory is used to assess the liquefaction safety factor F. SPT and F Vs To perform information fusion;

[0012] S4: Calculate the liquefaction probability interval [Bel(A),Pl(A)], where Bel(A) is the liquefaction confidence level and Pl(A) is the liquefaction likelihood level;

[0013] S5: Make a liquefaction decision based on the relationship between the liquefaction confidence level Bel(A) and the non-liquefaction confidence level Bel(B).

[0014] In a preferred embodiment of the present invention, the formula for calculating the liquefaction safety factor in step S2 is as follows:

[0015] ;

[0016] Where CSR is the cyclic stress ratio and CRR is the cyclic resistance ratio, the formula for calculating the cyclic stress ratio CSR is:

[0017] ;

[0018] In the formula, σ v For the total overburden stress, σ ' v For effective overburden stress, a max γ is the peak ground acceleration, g is the gravitational acceleration, and γ is the gravitational acceleration. d K is the shear stress reduction factor, MSF is the magnitude scaling factor, and K is the magnitude scaling factor. σ This is the overburden stress correction factor.

[0019] As a preferred embodiment of the present invention, the cycle resistance ratio (CRR) based on SPT-N data is... SPT Calculated using the following formula:

[0020]

[0021] Among them (N1) {60CS} To correct for the standard penetration test (SPT) blow count, the following formula is used:

[0022] ;

[0023] ;

[0024] ;

[0025] FC represents the percentage of fine particles (N1). 60 This represents the standard penetration test (SPT) blow count at the site.

[0026] As a preferred embodiment of the present invention, the circulation drag ratio (CRR) based on shear wave velocity Vs data is... Vs Calculated using the following formula:

[0027] ;

[0028] Where V s1 To correct for the shear wave velocity, the calculation formula is as follows:

[0029] ;

[0030] a and b are site calibration parameters, σ ' v For effective overburden stress, P a Standard atmospheric pressure.

[0031] In a preferred embodiment of the present invention, S3 specifically includes:

[0032] Support calculation: Calculate the liquefaction support, non-liquefaction support, and uncertainty using the following formulas:

[0033] ;

[0034] Basic probability assignment: Normalize the support to the basic probability assignment function:

[0035] ;

[0036] Evidence fusion: The Dempster combination rule is used for fusion of two evidence sources. The formula for calculating the conflict coefficient K is as follows:

[0037] ;

[0038] The formula for calculating the probability allocation after fusion is:

[0039] .

[0040] In a preferred embodiment of the present invention, in step S4, the formula for calculating the liquefaction probability interval is:

[0041] ;

[0042] The liquefaction probability interval is represented as [Bel(A),Pl(A)].

[0043] In a preferred embodiment of the present invention, in step S5, the liquefaction decision rule is as follows:

[0044] When Bel(A) > Bel(B), the soil is determined to be liquefied; otherwise, it is determined to be non-liquefied.

[0045] As a preferred embodiment of the present invention, it further includes a spatial variability processing step, which employs random field theory to process the spatial variability of soil parameters, specifically including:

[0046] Determine the horizontal and vertical correlation distances δ h and δ v ;

[0047] Establish a spatial correlation matrix R, with matrix elements R c,d The calculation formula is:

[0048]

[0049] The correlation matrix is ​​decomposed using the Cholesky decomposition method: R = LL T ;

[0050] Generate random fields that satisfy spatial correlation.

[0051] As a preferred embodiment of the present invention, the relevant distance δ h and δ v The likelihood function is determined using the maximum likelihood estimation method:

[0052]

[0053] Where D is the matrix of measured soil parameters at the borehole, and C D Let be the covariance matrix.

[0054] In a preferred embodiment of the present invention, a result verification step is further included, wherein the accuracy of the judgment result is verified by reserving borehole data, and the verification formula is as follows:

[0055] ;

[0056] Where L A (u α ) is the grid u α The accuracy of the liquefaction discrimination result, IndexStra(u α ) j Let be the indicator function for the j-th simulation.

[0057] Compared to traditional single-data-source methods, this invention integrates two independent evidence sources—the Standard Penetration Test (SPT-N) and the shear wave velocity Vs—using the Dempster-Shafer evidence theory, and has the following beneficial technical effects:

[0058] (1) Significantly improved discrimination accuracy: The complementary nature of dual evidence sources can compensate for the measurement errors and uncertainties caused by spatial variations in a single testing method;

[0059] (2) Probabilistic output: The probability interval formed by confidence Bel(A) and similarity Pl(A) provides the upper and lower bounds of the liquefaction probability, which facilitates engineering decision-making;

[0060] (3) Uncertainty quantification: The degree of conflict of evidence is reflected by the uncertainty m(θ), and a warning is given when the conflict between the two sources of evidence is large;

[0061] (4) Spatial variability processing: Combining random field theory and maximum likelihood estimation, it more realistically reflects the spatial distribution characteristics of soil parameters;

[0062] (5) Wide range of applications: It is particularly suitable for liquefaction risk assessment under complex geological conditions. Attached Figure Description

[0063] Figure 1 This is a flowchart illustrating a multi-source information fusion method for sand liquefaction discrimination based on evidence theory in this embodiment.

[0064] Figure 2 This is a schematic diagram showing how the liquefaction probability range changes with depth in this embodiment.

[0065] Figure 3 This is a schematic diagram comparing the security coefficients of dual evidence sources in this embodiment.

[0066] Figure 4 This is a schematic diagram illustrating the change in probability intervals after evidence fusion in this embodiment. Detailed Implementation

[0067] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0068] like Figure 1 As shown, a multi-source information fusion method for sand liquefaction discrimination based on Dempster-Shafer evidence theory is proposed. The method includes the following steps:

[0069] S1: Obtain the geotechnical history exploration dataset of the target area. This dataset includes dynamic parameters obtained through in-situ testing techniques, as well as static and physical state parameters obtained through laboratory geotechnical tests. The dynamic parameters include Standard Penetration Test (SPT-N) data and shear wave velocity data V. s ;

[0070] S2: Calculate the corresponding liquefaction safety factor F based on the standard penetration test data SPT-N and shear wave velocity data Vs, respectively. SPT and F Vs ;

[0071] S3: The Dempster-Shafer evidence theory is used to assess the liquefaction safety factor F. SPT and F Vs To perform information fusion; Figure 2 The figure illustrates the probability interval distribution of the liquefaction confidence level Bel(A) and the plausibility level Pl(A) along the depth direction of the target site after fusion using evidence theory. As can be seen from the figure, the width of the probability interval varies at different depths; a wider interval indicates a greater conflict and higher uncertainty between the two evidence sources (SPT-N and Vs) at that depth. Furthermore, the overall position of the interval (especially the magnitude of Bel(A)) shows a regular change with depth, with shallow soil often having a higher upper limit for liquefaction plausibility, indicating that shallow sand layers have a relatively greater risk of liquefaction under seismic action.

[0072] S4: Calculate the liquefaction probability interval [Bel(A),Pl(A)], where Bel(A) is the liquefaction confidence level and Pl(A) is the liquefaction likelihood level;

[0073] S5: Make a liquefaction decision based on the relationship between the liquefaction confidence level Bel(A) and the non-liquefaction confidence level Bel(B). For example... Figure 4 As shown, liquefaction decisions are made based on the relationship between Bel(A) and the non-liquefaction confidence level Bel(B). The figure illustrates the distribution of the liquefaction confidence level Bel(A), the non-liquefaction confidence level Bel(B), and the uncertainty m(θ) with depth after fusing F_SPT and F_Vs using Dempster-Shafer evidence theory. It can be seen from the figure that the boundary between the fused Bel(A) and Bel(B) curves is clearer, whereas previously... Figure 3 The depth segment that was difficult to determine due to dual-source conflict was clearly determined after fusion by the relationship between Bel(A) and Bel(B); at the same time, the segment with large uncertainty m(θ) was significantly marked, providing an intuitive basis for engineers to identify risk points that are "highly conflicting and require supplementary investigation".

[0074] In a preferred embodiment, the formula for calculating the liquefaction safety factor in step S2 is as follows:

[0075] ;

[0076] Where CSR is the cyclic stress ratio and CRR is the cyclic resistance ratio, the formula for calculating the cyclic stress ratio CSR is:

[0077] ;

[0078] In the formula, σ v For the total overburden stress, σ ' v For effective overburden stress, a max γ is the peak ground acceleration, g is the gravitational acceleration, and γ is the gravitational acceleration. d K is the shear stress reduction factor, MSF is the magnitude scaling factor, and K is the magnitude scaling factor. σ This is the overburden stress correction factor.

[0079] As a preferred embodiment, the cycle resistance ratio (CRR) based on SPT-N data is... SPT Calculated using the following formula:

[0080]

[0081] Among them (N1) {60CS} To correct for the standard penetration test (SPT) blow count, the following formula is used:

[0082] ;

[0083] ;

[0084] ;

[0085] FC represents the percentage of fine particles (N1). 60 This represents the standard penetration test (SPT) blow count at the site.

[0086] As a preferred embodiment, the circulation drag ratio (CRR) is based on shear wave velocity Vs data. Vs Calculated using the following formula:

[0087] ;

[0088] Where V s1 To correct for the shear wave velocity, the calculation formula is as follows:

[0089] ;

[0090] a and b are site calibration parameters, σ ' v For effective overburden stress, P a Standard atmospheric pressure.

[0091] like Figure 3The figure shows a comparison of safety factors from two sources of evidence. It compares the liquefaction safety factor FSPT calculated based on standard penetration test data with the liquefaction safety factor FVs calculated based on shear wave velocity data on the same depth axis. As can be seen from the figure, the safety factor curves calculated from the two sources of evidence do not overlap in most depth segments, and even show some deviation. When both FSPT and FVs are greater than 1, the conclusions from both sources consistently point to no liquefaction; when both are less than 1, they consistently point to liquefaction; and when one source of evidence is greater than 1 and the other is less than 1, a conflict of evidence arises. This is precisely the ambiguous area that traditional single-source data methods struggle to resolve.

[0092] As a preferred embodiment, S3 specifically includes:

[0093] Support calculation: Calculate the liquefaction support, non-liquefaction support, and uncertainty using the following formulas:

[0094] ;

[0095] Basic probability assignment: Normalize the support to the basic probability assignment function:

[0096] ;

[0097] Evidence fusion: The Dempster combination rule is used for fusion of two evidence sources. The formula for calculating the conflict coefficient K is as follows:

[0098] ;

[0099] The formula for calculating the probability allocation after fusion is:

[0100] .

[0101] In a preferred embodiment, in step S4, the formula for calculating the liquefaction probability interval is:

[0102] ;

[0103] The liquefaction probability interval is represented as [Bel(A),Pl(A)].

[0104] In a preferred embodiment of the present invention, in step S5, the liquefaction decision rule is as follows:

[0105] When Bel(A) > Bel(B), the soil is determined to be liquefied; otherwise, it is determined to be non-liquefied.

[0106] As a preferred embodiment, the method further includes a spatial variability processing step, which uses random field theory to process the spatial variability of soil parameters, specifically including:

[0107] Determine the horizontal and vertical correlation distances δ h and δ v ;

[0108] Establish a spatial correlation matrix R, with matrix elements R c,d The calculation formula is:

[0109]

[0110] The correlation matrix is ​​decomposed using the Cholesky decomposition method: R = LL T ;

[0111] Generate random fields that satisfy spatial correlation.

[0112] As a preferred embodiment, the relevant distance δ h and δ v The likelihood function is determined using the maximum likelihood estimation method:

[0113]

[0114] Where D is the matrix of measured soil parameters at the borehole, and C D Let be the covariance matrix.

[0115] As a preferred embodiment, the method further includes a result verification step, which verifies the accuracy of the judgment result using reserved borehole data. The verification formula is as follows:

[0116] ;

[0117] Where L A (u α ) is the grid u α The accuracy of the liquefaction discrimination result, IndexStra(u α ) j Let be the indicator function for the j-th simulation.

[0118] Preferably, the present invention may also employ the following alternatives:

[0119] (1) Substitution of sources of evidence: except for SPT-N and V s In addition, static cone penetration test (CPT) data can be introduced as a third source of evidence and fused using the multi-source Dempster combination rule;

[0120] (2) Support function replacement: Distance-based support construction methods, such as S A = 1 / (1+|F-1|), S B = 1 / (1+|F-1 / F|);

[0121] (3) Spatial variability substitution: In addition to random field theory, Bayesian update method or Kriging interpolation method can also be used to deal with the spatial variability of parameters;

[0122] (4) Substitution of fusion rule: When there is a large conflict of evidence, the Yager combination rule or the weighted average method can be used to replace the Dempster combination rule.

[0123] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for determining sand liquefaction based on multi-source information fusion using evidence theory, characterized in that, Includes the following steps: S1: Obtain a geotechnical historical exploration dataset of a target area site, the geotechnical historical exploration dataset containing dynamic parameters obtained by in-situ testing techniques and static mechanical parameters and physical state parameters obtained by indoor geotechnical tests; the dynamic parameters include standard penetration test data SPT-N and shear wave velocity data V s ; S2: Calculate the corresponding liquefaction safety factor F based on the standard penetration test data SPT-N and shear wave velocity data Vs, respectively. SPT and F Vs ; S3: The Dempster-Shafer evidence theory is used to assess the liquefaction safety factor F. SPT and F Vs To perform information fusion; S4: Calculate the liquefaction probability interval [Bel(A),Pl(A)], where Bel(A) is the liquefaction confidence level and Pl(A) is the liquefaction likelihood level; S5: Make a liquefaction decision based on the relationship between the liquefaction confidence level Bel(A) and the non-liquefaction confidence level Bel(B).

2. The method for determining sand liquefaction based on evidence theory using multi-source information fusion as described in claim 1, characterized in that, In S2, the formula for calculating the liquefaction safety factor is: ; Where CSR is the cyclic stress ratio and CRR is the cyclic resistance ratio, the formula for calculating the cyclic stress ratio CSR is: ; In the formula, σ v For the total overburden stress, σ ' v For effective overburden stress, a max γ is the peak ground acceleration, g is the gravitational acceleration, and γ is the gravitational acceleration. d K is the shear stress reduction factor, MSF is the magnitude scaling factor, and K is the magnitude scaling factor. σ This is the overburden stress correction factor.

3. The method for determining sand liquefaction based on evidence theory using multi-source information fusion as described in claim 2, characterized in that, Cyclic resistance ratio (CRR) based on SPT-N data SPT Calculated using the following formula: ; Among them (N1) {60CS} To correct for the standard penetration test (SPT) blow count, the following formula is used: ; ; ; FC represents the percentage of fine particles (N1). 60 This represents the standard penetration test (SPT) blow count at the site.

4. The method for determining sand liquefaction based on evidence theory using multi-source information fusion as described in claim 2, characterized in that, Based on shear wave velocity Vs data, the circulation drag ratio CRR Vs Calculated using the following formula: ; Where V s1 To correct for the shear wave velocity, the calculation formula is as follows: ; a and b are site calibration parameters, σ ' v For effective overburden stress, P a Standard atmospheric pressure.

5. The method for determining sand liquefaction based on evidence theory using multi-source information fusion as described in claim 1, characterized in that, S3 specifically includes: Support calculation: Calculate the liquefaction support, non-liquefaction support, and uncertainty using the following formulas: ; Basic probability assignment: Normalize the support to the basic probability assignment function: ; Evidence fusion: The Dempster combination rule is used for fusion of two evidence sources. The formula for calculating the conflict coefficient K is as follows: ; The formula for calculating the probability allocation after fusion is: 。 6. The method for determining sand liquefaction based on evidence theory using multi-source information fusion as described in claim 1, characterized in that, In S4, the formula for calculating the liquefaction probability interval is: ; The liquefaction probability interval is represented as [Bel(A),Pl(A)].

7. The method for determining sand liquefaction based on evidence theory using multi-source information fusion as described in claim 1, characterized in that, In S5, the liquefaction decision rule is: When Bel(A) > Bel(B), the soil is determined to be liquefied; otherwise, it is determined to be non-liquefied.

8. The method for determining sand liquefaction based on evidence theory using multi-source information fusion as described in claim 1, characterized in that, It also includes a spatial variability processing step, which uses random field theory to handle the spatial variability of soil parameters, specifically including: Determine the horizontal and vertical correlation distances δ h and δ v ; Establish a spatial correlation matrix R, with matrix elements R c,d The calculation formula is: ; The correlation matrix is ​​decomposed using the Cholesky decomposition method: R = LL T ; Generate random fields that satisfy spatial correlation.

9. The method for determining sand liquefaction based on evidence theory using multi-source information fusion as described in claim 8, characterized in that, The relevant distance δ h and δ v The likelihood function is determined using the maximum likelihood estimation method: ; Where D is the matrix of measured soil parameters at the borehole, and C D Let be the covariance matrix.

10. The method for determining sand liquefaction based on evidence theory using multi-source information fusion as described in claim 1, characterized in that, It also includes a result verification step, which verifies the accuracy of the judgment result using reserved borehole data. The verification formula is: ; Where L A (u α ) is the grid u α The accuracy of the liquefaction discrimination result, IndexStra(u α ) j Let be the indicator function for the j-th simulation.