Ground stress direction measurement system based on longitude and latitude strain lines

By combining a two-dimensional orthogonal strain acquisition array with ground-penetrating radar, and employing dynamic correction and weighted fusion techniques, the problems of insufficient accuracy and anti-interference capability of geostress measurement systems in complex rock mass environments have been solved, enabling high-precision stress direction monitoring and risk early warning.

CN121185478AActive Publication Date: 2025-12-23NAT INST OF NATURAL HAZARDS MINISTRY OF EMERGENCY MANAGEMENT OF CHINA
View PDF 10 Cites 0 Cited by

Patent Information

Application Number
CN202511083999.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-08-04
Publication Date
2025-12-23
Estimated Expiration
2045-08-04

AI Technical Summary

Technical Problem

Existing geostress measurement systems cannot effectively identify the heterogeneity of rock mass structures in complex geological environments, resulting in insufficient accuracy in inverting the principal stress direction. Furthermore, they lack the ability to identify differences in data reliability, which can easily lead to systematic biases.

Method used

A two-dimensional orthogonal strain acquisition array, time-series synchronous control, interference adaptive filtering, structural adaptive compensation and dynamic weighted fusion technology are adopted. Combined with the fracture distribution model obtained by the ground-penetrating radar module, the stress direction is dynamically corrected and weighted fused to construct a high-precision geostress tensor.

Benefits of technology

It improves the accuracy of main direction identification and anti-abnormal interference capability of the geostress measurement system in complex rock mass environments, significantly reduces the misjudgment rate, and realizes the monitoring of risk evolution sequence of stress concentration areas.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121185478A_ABST
    Figure CN121185478A_ABST
Patent Text Reader

Abstract

The invention discloses a crustal stress direction measurement system based on longitude and latitude strain lines, and belongs to the technical field of crustal stress measurement. Comprising a two-dimensional orthogonal strain acquisition array module; the time sequence synchronous control module is used for controlling each group of strain sensors to perform synchronous sampling based on a unified time reference; an interference adaptive filtering module; a crustal stress tensor solving module; a structure adaptive compensation module; a dynamic weighted fusion module; a risk trend deduction module; the edge calculation module is connected with the remote communication module. According to the system, the fiber bragg grating sensing array in which the longitude and the latitude are orthogonally arranged is adopted, a symmetrical strain tensor is constructed, eigenvalue decomposition is carried out, so that a principal stress direction angle and a principal strain value are obtained, and compared with a single-point strain gauge or a traditional trepanning method, continuous principal direction distribution reconstruction can be realized in a larger range, and the precision of the system is improved. The method has higher spatial resolution and tensor integrity.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of geostress measurement technology, specifically relating to a geostress direction measurement system based on latitude and longitude strain lines. Background Technology

[0002] In-situ stress is a fundamental physical quantity present in rock masses and a crucial controlling factor affecting the stability of underground engineering, rock mass failure mechanisms, and tectonic evolution. Accurately determining the magnitude of the in-situ stress tensor and the direction of the principal stresses has critical engineering guidance value for deep-earth engineering projects such as tunnel excavation, deep well mining, and energy storage design.

[0003] With the development of distributed fiber optic sensing technology, high-precision strain field measurement using orthogonally deployed longitudinal and latitudinal strain fiber arrays is gradually becoming a trend. This type of method can acquire two-dimensional strain tensor information and derive the principal stress directions through tensor principal value calculation. However, most existing methods only apply to ideal homogeneous soil conditions and fail to consider the interference of fracture structures that are commonly present in rock masses, resulting in insufficient accuracy of the stress principal direction inversion results in complex geological environments.

[0004] On the other hand, most fiber optic strain measurement systems currently only average and fuse the strain values ​​of each sensing node, without establishing a weighting mechanism based on quality indicators such as node signal-to-noise ratio, stability, and deployment density. They lack the ability to identify differences in data reliability, which can easily lead to the overall stress tensor calculation process being affected by local anomalies and resulting in systematic deviations.

[0005] Although existing studies have attempted to use structural sensing methods such as ground-penetrating radar (GPR) for rock mass structure identification, there is still a lack of a fusion path in the current geostress measurement system to incorporate GPR fracture model data into stress direction correction, meaning that it is impossible to dynamically correct and compensate for structural heterogeneity in strain tensor calculation. Summary of the Invention

[0006] The present invention is as follows:

[0007] A geostress direction measurement system based on latitude and longitude strain lines includes:

[0008] A two-dimensional orthogonal strain acquisition array module is composed of a group of latitude strain sensors arranged along the X-axis and a group of longitude strain sensors arranged along the Y-axis. Each group consists of at least two sensors at different depth levels, forming a three-dimensional sampling grid.

[0009] The timing synchronization control module is used to control the synchronous sampling of each group of strain sensors based on a unified time reference;

[0010] The interference adaptive filtering module includes a wavelet packet decomposition unit, an interference identification submodule, and a filtering reconstruction submodule;

[0011] The geostress tensor solving module is used to construct a two-dimensional strain tensor matrix based on latitude and longitude strain values ​​and solve for its principal direction angles.

[0012] The structural adaptive compensation module is used to construct an anisotropic correction tensor using the crack distribution model obtained from the external GPR and apply it to the original stress direction.

[0013] The dynamic weighted fusion module constructs a weight vector based on the signal-to-noise ratio, temporal stability, and spatial distribution density of each measurement point and applies it to tensor calculation.

[0014] The risk trend projection module is used to generate risk evolution sequences for stress concentration areas based on the abrupt change rate of geostress direction.

[0015] The edge computing module and remote communication module are integrated into the field node box, and have data preprocessing, local direction calculation and 5G transmission capabilities.

[0016] Preferably, both the latitude strain sensor and the longitude strain sensor are fiber Bragg grating strain gauges, specifically including a Bragg grating region, a reflective mirror encapsulation structure, and a flexible mechanical coupling sheet. Each strain gauge is connected in series on a single optical fiber via wavelength division multiplexing and is evenly distributed on the monitoring surface at 0.5-meter intervals. Each sensor is embedded into the rock surface at a depth of 0.8 meters through a fixed-depth borehole, with a borehole depth tolerance not exceeding ±0.1 meters. A flexible filling fixative is used to ensure tight coupling between the fiber Bragg grating sensor head and the rock mass. The reflection wavelength range of each strain gauge is set to 1525-1565nm, and the center wavelength drift sensitivity is greater than 1.1pm / με. The optical cable sheath is covered with a double-layer flexible silicone sealant and a PTFE geomembrane layer to ensure long-term stable operation in a geothermal environment ranging from -20℃ to 80℃.

[0017] Preferably, the interference adaptive filtering module adopts a multi-stage combined filtering structure, including:

[0018] The wavelet packet decomposition unit performs 5-level full-band signal decomposition based on the Daubechies-8th order wavelet function, with a signal processing frame length of 8192 sampling points.

[0019] The interference identification submodule has a built-in feature library that includes spectral templates for at least 12 types of seismic disturbances, mechanical disturbances, and electromagnetic interferences, and automatically filters out anomalous components using a combination of sparse matching pursuit and Hilbert amplitude analysis.

[0020] The filtering and reconstruction submodule uses a Kalman prediction-wavelet packet reconstruction fusion algorithm to compensate and backfill the identified effective strain signals, and then superimposes an adaptive Savitzky-Golay window on the reconstruction results to perform second-order smoothing on the edges of the step signal.

[0021] Preferably, the geostress tensor solving module is based on the fused strain tensor ε w Constructing a two-dimensional symmetric strain tensor:

[0022]

[0023] Where: ε xx ε yy ε represents the average linear strain value along the parallels of latitude and longitude, respectively; xy The shear strain component was calculated using mutual inductance sensors spaced no more than 0.2m apart in orthogonal directions.

[0024]

[0025] Principal stress direction angle θ raw With principal strain value ε 1,2 The solution is obtained by using the eigenvalues ​​and eigenvectors of the tensor, respectively, and the formula is as follows:

[0026]

[0027] Preferably, the structural adaptive compensation module is used to adjust the principal stress direction angle θ. raw The fracture structure coupling correction is performed. The fracture distribution model is acquired by the system-integrated ground-penetrating radar (GPR) module. The GPR module includes: a dual-polarized antenna with an operating frequency band of 400-800MHz; a vehicle-mounted towed platform or a wall-mounted scanning arm; the radar scan data is processed by background removal, gain compensation and Kirchhoff inversion to form a two-dimensional B-scan profile.

[0028] The image was processed using edge detection and the Hough transform algorithm to extract the main fracture strike angle φ. i With density D i The structural compensation angle Δθ is generated using the following formula:

[0029] Corrected principal stress direction angle θ corr The calculation formula is: θ corr =θ raw +Δθ.

[0030] Preferably, the strain tensor ε is fused. w The generation process is completed by a dynamic weighted fusion module, which uses the strain tensor ε collected by N distributed sensing nodes. i The following weighted method is used for synthesis:

[0031] Where the weighting coefficient w i Constructed based on node quality factors, specifically:

[0032]

[0033] Among them: SNR i STAB represents the signal-to-noise ratio of the i-th node; i DENS represents the reciprocal of the standard deviation of the strain at the node over the most recent 5 minutes. i This represents the average density of nodes within their respective sub-regions; α, β, and γ are empirical weighting coefficients that satisfy α + β + γ = 1.

[0034] Preferably, the risk trend deduction module uses the principal stress direction angle θ obtained by the structural adaptive compensation module. corr Based on this, combined with the fused strain tensor ε output by the dynamic weighted fusion module w Based on the principal strain value sequence, a geostress mutation rate monitoring model is constructed, which includes the following processing steps:

[0035] Step 1: Convert the continuous principal direction angles θ corr (t) Perform time series differencing using a 1-minute sliding window and calculate the directional abrupt change rate: Δθ(t)=θ corr (t)-θ corr (t-Δt);

[0036] Step 2: Simultaneously extract the principal strain feature values ​​ε1(t) and ε2(t), and construct a two-dimensional risk feature vector:

[0037] Step 3: Input the feature vector into the Support Vector Machine (SVM) classification model and compare it with the training set samples to determine whether the current area is in a low, medium, or high risk state.

[0038] Step 4: Record the risk level results within a continuous time period as a risk evolution sequence R(t), and generate a stress concentration trend map and early warning signal for the current monitoring area based on the spatial coordinate correspondence.

[0039] The feature vector dimension, sliding window size, and classification threshold of the SVM model can be set by the user or adjusted adaptively by the system.

[0040] Preferably, the edge computing module is built on the NVIDIA Jetson Orin Nano platform, which embeds a direction solving module, a filtering module and a risk model evaluation module, and has a built-in dual-channel configuration of LoRa wireless communication module and 5G module. When the signal attenuation is greater than 30dB, the communication channel is automatically switched to ensure data link redundancy.

[0041] Preferably, the system is used in tunnels crossing fault zones. It forms a long-axis control network by deploying a multi-layer orthogonal strain array every 20 meters, and continuously calculates the principal stress variation sequence along the tunnel axis through tensor direction field to determine the specific locations of weak zones, tension concentration zones and potential deformation fault zones.

[0042] Compared with the prior art, the advantages of the present invention are as follows:

[0043] (1) This system uses a fiber optic grating sensing array orthogonally arranged with meridians and parallels to construct a symmetrical strain tensor and perform eigenvalue decomposition, thereby obtaining the principal stress direction angle and principal strain value. Compared with single-point strain gauges or traditional hole-punch method, this invention can realize continuous principal direction distribution reconstruction over a larger range, and has higher spatial resolution and tensor integrity.

[0044] (2) The system quantifies parameters such as signal-to-noise ratio, time stability, and layout density of each strain acquisition point to construct a fusion weight vector, thereby achieving a multi-point weighted average of the initial strain tensor. Compared with traditional averaging or filtering methods, this invention improves the reliability of tensor calculation and its ability to resist abnormal interference, and significantly reduces the stress direction misjudgment rate.

[0045] (3) This system incorporates a ground-penetrating radar (GPR) module to acquire information on the spatial distribution, orientation, and density of fractures within the monitoring area. It also compensates for and corrects the original principal stress direction angles by constructing an anisotropic structure correction tensor. This approach solves the problem of existing measurement systems being unable to detect the heterogeneity of the medium structure, and provides higher principal direction identification accuracy in complex rock mass environments. Attached Figure Description

[0046] Figure 1 This is a block diagram of a geostress direction measurement system based on latitude and longitude strain lines. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be explained and described below. However, the following embodiments are only preferred embodiments of the present invention and not all of them. Other embodiments obtained by those skilled in the art based on the embodiments in the implementation methods without creative effort are all within the protection scope of the present invention.

[0048] Example 1

[0049] A geostress direction measurement system based on latitude and longitude strain lines includes:

[0050] A two-dimensional orthogonal strain acquisition array module is composed of a group of latitude strain sensors arranged along the X-axis and a group of longitude strain sensors arranged along the Y-axis. Each group consists of at least two sensors at different depth levels, forming a three-dimensional sampling grid.

[0051] The timing synchronization control module is used to control the synchronous sampling of each group of strain sensors based on a unified time reference;

[0052] The interference adaptive filtering module includes a wavelet packet decomposition unit, an interference identification submodule, and a filtering reconstruction submodule;

[0053] The geostress tensor solving module is used to construct a two-dimensional strain tensor matrix based on latitude and longitude strain values ​​and solve for its principal direction angles.

[0054] The structural adaptive compensation module is used to construct an anisotropic correction tensor using the crack distribution model obtained from the external GPR and apply it to the original stress direction.

[0055] The dynamic weighted fusion module constructs a weight vector based on the signal-to-noise ratio, temporal stability, and spatial distribution density of each measurement point and applies it to tensor calculation.

[0056] The risk trend projection module is used to generate risk evolution sequences for stress concentration areas based on the abrupt change rate of geostress direction.

[0057] The edge computing module and remote communication module are integrated into the field node box, and have data preprocessing, local direction calculation and 5G transmission capabilities.

[0058] Both the latitude and longitude strain sensors are fiber Bragg grating strain gauges, specifically comprising a Bragg grating region, a reflective mirror encapsulation structure, and a flexible mechanical coupling sheet. Each strain gauge is connected in series on a single optical fiber via wavelength division multiplexing and is evenly distributed on the monitoring surface at 0.5-meter intervals. Each sensor is embedded into the rock surface at a depth of 0.8 meters through fixed-depth drilling, with a drilling depth tolerance not exceeding ±0.1 meters. Flexible filling and fixing agents are used to ensure tight coupling between the fiber Bragg grating sensor head and the rock mass. The reflection wavelength range of each strain gauge is set to 1525-1565nm, and the center wavelength drift sensitivity is greater than 1.1pm / με. The optical cable sheath is covered with a double-layer flexible silicone sealant and a PTFE geomembrane layer to ensure long-term stable operation in a geothermal environment ranging from -20℃ to 80℃.

[0059] The interference adaptive filtering module adopts a multi-level combined filtering structure, including:

[0060] The wavelet packet decomposition unit performs 5-level full-band signal decomposition based on the Daubechies-8th order wavelet function, with a signal processing frame length of 8192 sampling points.

[0061] The interference identification submodule has a built-in feature library that includes spectral templates for at least 12 types of seismic disturbances, mechanical disturbances, and electromagnetic interferences, and automatically filters out anomalous components using a combination of sparse matching pursuit and Hilbert amplitude analysis.

[0062] The filtering and reconstruction submodule uses a Kalman prediction-wavelet packet reconstruction fusion algorithm to compensate and backfill the identified effective strain signals, and then superimposes an adaptive Savitzky-Golay window on the reconstruction results to perform second-order smoothing on the edges of the step signal.

[0063] The geostress tensor solution module is based on the fused strain tensor ε. w Constructing a two-dimensional symmetric strain tensor:

[0064]

[0065] Where: ε xx ε yy ε represents the average linear strain value along the parallels of latitude and longitude, respectively; xy The shear strain component was calculated using mutual inductance sensors spaced no more than 0.2m apart in orthogonal directions.

[0066]

[0067] Principal stress direction angle θ raw With principal strain value ε 1,2 The solution is obtained by using the eigenvalues ​​and eigenvectors of the tensor, respectively, and the formula is as follows:

[0068]

[0069] The structural adaptive compensation module is used to adjust the principal stress direction angle θ. raw The fracture structure coupling correction is performed. The fracture distribution model is acquired by the system-integrated ground-penetrating radar (GPR) module. The GPR module includes: a dual-polarized antenna with an operating frequency band of 400-800MHz; a vehicle-mounted towed platform or a wall-mounted scanning arm; the radar scan data is processed by background removal, gain compensation and Kirchhoff inversion to form a two-dimensional B-scan profile.

[0070] The image was processed using edge detection and the Hough transform algorithm to extract the main fracture strike angle φ. i With density D i The structural compensation angle Δθ is generated using the following formula:

[0071] Corrected principal stress direction angle θ corr The calculation formula is: θ corr =θ raw +Δθ.

[0072] Among them, the fused strain tensor ε w The generation process is completed by a dynamic weighted fusion module, which uses the strain tensor ε collected by N distributed sensing nodes.i The following weighted method is used for synthesis:

[0073] Where the weighting coefficient w i Constructed based on node quality factors, specifically:

[0074]

[0075] Among them: SNR i STAB represents the signal-to-noise ratio of the i-th node; i DENS represents the reciprocal of the standard deviation of the strain at the node over the most recent 5 minutes. i This represents the average density of nodes within their respective sub-regions; α, β, and γ are empirical weighting coefficients that satisfy α + β + γ = 1.

[0076] The risk trend deduction module uses the principal stress direction angle θ obtained after correction by the structural adaptive compensation module. corr Based on this, combined with the fused strain tensor ε output by the dynamic weighted fusion module w Based on the principal strain value sequence, a geostress mutation rate monitoring model is constructed, which includes the following processing steps:

[0077] Step 1: Convert the continuous principal direction angles θ corr (t) Perform time series differencing using a 1-minute sliding window and calculate the directional abrupt change rate: Δθ(t)=θ corr (t)-θ corr (t-Δt);

[0078] Step 2: Simultaneously extract the principal strain feature values ​​ε1(t) and ε2(t), and construct a two-dimensional risk feature vector:

[0079] Step 3: Input the feature vector into the Support Vector Machine (SVM) classification model and compare it with the training set samples to determine whether the current area is in a low, medium, or high risk state.

[0080] Step 4: Record the risk level results within a continuous time period as a risk evolution sequence R(t), and generate a stress concentration trend map and early warning signal for the current monitoring area based on the spatial coordinate correspondence.

[0081] The feature vector dimension, sliding window size, and classification threshold of the SVM model can be set by the user or adjusted adaptively by the system.

[0082] The edge computing module is built on the NVIDIA Jetson Orin Nano platform. It embeds a direction solving module, a filtering module, and a risk model evaluation module. It also has a built-in dual-channel configuration of LoRa wireless communication module and 5G module. When the signal attenuation is greater than 30dB, it automatically switches the communication channel to ensure data link redundancy.

[0083] The system is used in tunnels crossing fault zones. It forms a long-axis control network by deploying a multi-layer orthogonal strain array every 20 meters, and continuously calculates the principal stress variation sequence along the tunnel axis through tensor direction field to determine the specific locations of weak zones, tension concentration zones and potential deformation fault zones.

[0084] Example 2

[0085] This embodiment provides a geostress direction measurement system based on latitude and longitude strain lines, which is suitable for initial stress detection and risk zoning identification in an underground tunnel excavation project. The installation area is the rock mass monitoring surface within 20m in front of the tunnel face.

[0086] The system consists of the following modules:

[0087] (1) Deployment of two-dimensional orthogonal strain acquisition array

[0088] A two-dimensional orthogonal acquisition array consisting of fiber optic strain sensors was deployed on the rock surface of the monitoring area, with a spacing of 0.25m × 0.25m. The longitude and latitude lines were 10m and 8m long, respectively, with a total of 128 sensing nodes installed. Each node can acquire axial strain and 45° shear strain, with a sampling frequency of 1Hz and an error not exceeding ±5με.

[0089] (2) Raw strain data processing and dynamic weighted fusion

[0090] The original strain values ​​collected After low-pass filtering and noise reduction, the signal is input to the dynamic weighted fusion module. A weighting factor w is constructed based on the signal-to-noise ratio (SNR), strain stability (STAB), and regional deployment density (DENS) of each node over the past 5 minutes. i The tensor contribution values ​​of each node are then merged:

[0091]

[0092] The fused two-dimensional strain tensor ε is obtained w The principal strain values ​​ε1 and ε2 are solved from their eigenvalues, and the preliminary principal stress direction angle θ is solved from their eigenvectors. raw .

[0093] (3) Crack distribution model acquisition and structural compensation tensor generation

[0094] A GPR scanning vehicle was deployed simultaneously in the monitoring area, and a mobile inversion imaging system using an 800MHz dual-polarized antenna was employed to extract the reflection features of the main fracture surface within a 5m depth underground. After edge detection and Hough transform processing, the fracture strike angle distribution φ was extracted. i and the corresponding fracture density D i Input the above data into the structural compensation module to construct the crack compensation tensor and calculate the principal direction correction angle:

[0095]

[0096] Thus, the corrected principal stress direction angle θ is obtained. corr =θ raw +Δθ.

[0097] (4) Risk trend projection and concentration area identification

[0098] For the corrected direction angle sequence θ corr (t) Perform time series moving difference, calculate the mutation rate Δθ(t), and simultaneously calculate the first derivative of ε1(t) to construct a risk feature vector:

[0099]

[0100] It inputs a risk classification model based on Support Vector Machine (SVM), outputs risk level labels (low, medium, high), marks stress concentration risk areas, and forms an evolution sequence.

[0101] (5) Results Output and Visualization

[0102] The system generates risk level heatmaps, direction field vector maps, and stress mutation trend maps on the server side, and uploads them to a remote platform for engineering technicians to refer to. If the risk level is determined to be "high" within a continuous time window, the system will trigger an audible and visual alarm and push an alert to the mobile terminal.

[0103] The system shown in this embodiment has a deployment cycle of no more than 24 hours and can achieve continuous 7×24-hour monitoring of key areas. It has been applied and verified in multiple heterogeneous rock strata and fracture zone crossing sections. The identification accuracy is better than ±8° and the sensitivity of direction change detection reaches 0.5° / min. It can be effectively used for surrounding rock support adjustment and early prediction of high-risk tunneling sections.

[0104] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A geostress direction measurement system based on latitude and longitude strain lines, characterized in that, include: A two-dimensional orthogonal strain acquisition array module is composed of a group of latitude strain sensors arranged along the X-axis and a group of longitude strain sensors arranged along the Y-axis. Each group consists of at least two sensors at different depth levels, forming a three-dimensional sampling grid. The timing synchronization control module is used to control the synchronous sampling of each group of strain sensors based on a unified time reference; The interference adaptive filtering module includes a wavelet packet decomposition unit, an interference identification submodule, and a filtering reconstruction submodule; The geostress tensor solving module is used to construct a two-dimensional strain tensor matrix based on latitude and longitude strain values ​​and solve for its principal direction angles. The structural adaptive compensation module is used to construct an anisotropic correction tensor using the crack distribution model obtained from the external GPR and apply it to the original stress direction. The dynamic weighted fusion module constructs a weight vector based on the signal-to-noise ratio, temporal stability, and spatial distribution density of each measurement point and applies it to tensor calculation. The risk trend projection module is used to generate risk evolution sequences for stress concentration areas based on the abrupt change rate of geostress direction. The edge computing module and remote communication module are integrated into the field node box, and have data preprocessing, local direction calculation and 5G transmission capabilities.

2. The geostress direction measurement system based on latitude and longitude strain lines according to claim 1, characterized in that, Both the latitude and longitude strain sensors are fiber Bragg grating strain gauges, specifically comprising a Bragg grating region, a reflective mirror encapsulation structure, and a flexible mechanical coupling sheet. The fiber Bragg grating strain gauges are connected in series on a single optical fiber via wavelength division multiplexing and are evenly distributed on the monitoring surface at 0.5-meter intervals. Each sensor is embedded into the rock surface at a depth of 0.8 meters through fixed-depth drilling, with a drilling depth tolerance not exceeding ±0.1 meters. Flexible filling and fixing agents are used to ensure tight coupling between the fiber Bragg grating sensor head and the rock mass. The reflection wavelength range of each strain gauge is set to 1525-1565nm, and the center wavelength drift sensitivity is greater than 1.1pm / με. The optical cable sheath is covered with a double-layer flexible silicone sealant and a PTFE geomembrane layer to ensure long-term stable operation in a geothermal environment ranging from -20℃ to 80℃.

3. The geostress direction measurement system based on latitude and longitude strain lines according to claim 1, characterized in that, The interference adaptive filtering module adopts a multi-level combined filtering structure, including: The wavelet packet decomposition unit performs 5-level full-band signal decomposition based on the Daubechies-8th order wavelet function, with a signal processing frame length of 8192 sampling points. The interference identification submodule has a built-in feature library that includes spectral templates for at least 12 types of seismic disturbances, mechanical disturbances, and electromagnetic interferences, and automatically filters out anomalous components using a combination of sparse matching pursuit and Hilbert amplitude analysis. The filtering and reconstruction submodule uses a Kalman prediction-wavelet packet reconstruction fusion algorithm to compensate and backfill the identified effective strain signals, and then superimposes an adaptive Savitzky-Golay window on the reconstruction results to perform second-order smoothing on the edges of the step signal.

4. The geostress direction measurement system based on latitude and longitude strain lines according to claim 1, characterized in that, The geostress tensor solver module is based on the fused strain tensor ε w Constructing a two-dimensional symmetric strain tensor: Where: ε xx ε yy ε represents the average linear strain value along the parallels of latitude and longitude, respectively; xy The shear strain component was calculated using mutual inductance sensors spaced no more than 0.2m apart in orthogonal directions. Principal stress direction angle θ raw With principal strain value ε 1,2 The solution is obtained by using the eigenvalues ​​and eigenvectors of the tensor, respectively, and the formula is as follows:

5. The geostress direction measurement system based on latitude and longitude strain lines according to claim 4, characterized in that, The structural adaptive compensation module is used to adjust the principal stress direction angle θ. raw The fracture structure coupling correction is performed. The fracture distribution model is acquired by the system-integrated ground-penetrating radar (GPR) module. The GPR module includes: a dual-polarized antenna with an operating frequency band of 400-800MHz; a vehicle-mounted towed platform or a wall-mounted scanning arm; the radar scan data is processed by background removal, gain compensation and Kirchhoff inversion to form a two-dimensional B-scan profile. The image was processed using edge detection and the Hough transform algorithm to extract the main fracture strike angle φ. i With density D i The structural compensation angle Δθ is generated using the following formula: Corrected principal stress direction angle θ corr The calculation formula is: θ corr =θ raw +Δθ.

6. The geostress direction measurement system based on latitude and longitude strain lines according to claim 4, characterized in that, Fusion strain tensor ε w The generation process is completed by a dynamic weighted fusion module, which uses the strain tensor ε collected by N distributed sensing nodes. i The following weighted method is used for synthesis: Where the weighting coefficient w i Constructed based on node quality factors, specifically: Among them: SNR i STAB represents the signal-to-noise ratio of the i-th node; i DENS represents the reciprocal of the standard deviation of the strain at the node over the most recent 5 minutes. i This represents the average density of nodes within their respective sub-regions; α, β, and γ are empirical weighting coefficients that satisfy α + β + γ = 1.

7. The geostress direction measurement system based on latitude and longitude strain lines according to claim 5, characterized in that, The risk trend projection module uses the principal stress direction angle θ, which has been corrected by the structural adaptive compensation module. corr Based on this, combined with the fused strain tensor ε output by the dynamic weighted fusion module w Based on the principal strain value sequence, a geostress mutation rate monitoring model is constructed, which includes the following processing steps: Step 1: Convert the continuous principal direction angles θ corr (t) Perform time series differencing using a 1-minute sliding window and calculate the directional abrupt change rate: Δθ(t)=θ corr (t)-θ corr (t-Δt); Step 2: Simultaneously extract the principal strain feature values ​​ε1(t) and ε2(t), and construct a two-dimensional risk feature vector: Step 3: Input the feature vector into the Support Vector Machine (SVM) classification model and compare it with the training set samples to determine whether the current area is in a low, medium, or high risk state. Step 4: Record the risk level results within a continuous time period as a risk evolution sequence R(t), and generate a stress concentration trend map and early warning signal for the current monitoring area based on the spatial coordinate correspondence. The feature vector dimension, sliding window size, and classification threshold of the SVM model can be set by the user or adjusted adaptively by the system.

8. The geostress direction measurement system based on latitude and longitude strain lines according to claim 1, characterized in that, The edge computing module is built on the NVIDIA Jetson Orin Nano platform. It embeds a direction solving module, a filtering module, and a risk model evaluation module. It also has a built-in dual-channel configuration of LoRa wireless communication module and 5G module. When the signal attenuation is greater than 30dB, it automatically switches the communication channel to ensure data link redundancy.

9. A geostress direction measurement system based on latitude and longitude strain lines according to claim 1, characterized in that, The system is used in tunnels crossing fault zones. It forms a long-axis control network by deploying a multi-layer orthogonal strain array every 20 meters, and continuously calculates the principal stress variation sequence along the tunnel axis through tensor direction field to determine the specific locations of weak zones, tension concentration zones and potential deformation fault zones.

Citation Information

Patent Citations

  • Determination method for magnitude and direction of crustal stress

    CN104390733A

  • Three-dimensional disturbance stress testing method for stope surrounding rock

    CN107529615A

  • Antholith stress measuring method based on underground directional long borehole hydraulic fracturing method

    CN115493737A

  • Ground stress real-time monitoring method based on multi-source data fusion

    CN119290208A

  • Tunnel multi-source fusion dynamic twin surrounding rock intelligent prediction and control method and system

    CN120087772A