Roadway surrounding rock deformation monitoring method based on optical fiber sensor

Abstract

The invention discloses a roadway surrounding rock deformation monitoring method based on an optical fiber sensor, and the method comprises the steps: laying on the surface of roadway surrounding rock through employing a distributed optical fiber sensing technology, and obtaining original optical signal data; processing backscattered light data in the original optical signal data through a Brillouin optical time domain analysis algorithm, extracting frequency deviation of an optical signal, and obtaining strain distribution data; according to the strain distribution data, in combination with a pre-established surrounding rock mechanical model, performing initial inversion calculation by adopting a finite element method to obtain initial three-dimensional deformation distribution; on the basis of the initial three-dimensional deformation distribution, updating dynamic boundary conditions of the surrounding rock mechanical model by acquiring real-time geological parameter data to obtain an optimized mechanical model; carrying out inversion calculation on the strain distribution data again by adopting an optimized mechanical model to generate final three-dimensional deformation distribution; and based on the final three-dimensional deformation distribution, performing comparison judgment through a preset deformation threshold, determining the stability state of the surrounding rock, and generating an early warning signal.

Description

Technical Field

[0001] The present invention belongs to the technical field of surrounding rock monitoring, and in particular relates to a tunnel surrounding rock deformation monitoring method based on optical fiber sensors. Background Art

[0002] Stability monitoring of tunnel surrounding rock is a core area of underground engineering safety management, directly related to personnel safety and project stability. With the expansion of underground engineering scale and the increase in complex geological environments, real-time and accurate understanding of surrounding rock deformation has become the key to ensuring construction safety. However, existing monitoring methods have significant limitations in practical applications. Traditional technologies rely heavily on point sensors with limited coverage, making it difficult to fully reflect the overall deformation characteristics of tunnel surrounding rock. In addition, the real-time and accuracy of data acquisition are often limited by environmental interference, especially under complex geological conditions, making it difficult to accurately capture the dynamic changes of deep surrounding rock. These limitations have led to insufficient early warning capabilities of the monitoring system for potential risks, increasing safety hazards. Summary of the Invention

[0003] To solve the above technical problems, the present invention proposes a tunnel surrounding rock deformation monitoring method based on optical fiber sensors, which can realize real-time and high-precision monitoring of tunnel surrounding rock deformation, providing a strong guarantee for mine safety production.

[0004] To achieve the above objectives, the present invention provides a method for monitoring deformation of surrounding rock in a tunnel based on an optical fiber sensor, comprising:

[0005] Distributed optical fiber sensing technology is deployed on the surface of the tunnel surrounding rock to obtain raw optical signal data;

[0006] Processing the backscattered light data in the original optical signal data by using a Brillouin optical time domain analysis algorithm, extracting the frequency offset of the optical signal, and obtaining strain distribution data;

[0007] Based on the strain distribution data and the pre-established surrounding rock mechanical model, the finite element method is used to perform initial inversion calculations to obtain the preliminary three-dimensional deformation distribution;

[0008] Based on the preliminary three-dimensional deformation distribution, the dynamic boundary conditions of the surrounding rock mechanical model are updated by acquiring real-time geological parameter data to obtain an optimized mechanical model;

[0009] The strain distribution data is inverted again using the optimized mechanical model to generate the final three-dimensional deformation distribution;

[0010] Based on the final three-dimensional deformation distribution, a comparison and judgment is made through the preset deformation threshold to determine the stability status of the surrounding rock and generate an early warning signal.

[0011] Optionally, distributed optical fiber sensing technology is deployed on the surface of the tunnel surrounding rock to obtain raw optical signal data including:

[0012] A sensor network is deployed on the surface of the tunnel surrounding rock using distributed optical fiber sensing technology to obtain initial optical signal data containing continuous backscattered light signals.

[0013] The wavelet transform algorithm is used to perform denoising on the initial optical signal data to obtain the original optical signal data.

[0014] Optionally, processing the backscattered light data in the original optical signal data by using a Brillouin optical time domain analysis algorithm to extract the frequency offset of the optical signal and obtain the strain distribution data includes:

[0015] By collecting and preprocessing the optical signal data, the backscattered signal is separated from the original data to obtain the initial scattered signal data;

[0016] The Brillouin optical time domain analysis method is used to perform time domain processing on the initial scattered signal data, extract characteristic information related to the frequency offset, and obtain the frequency offset feature set.

[0017] Based on the frequency offset feature set, a preset offset calculation model is used to perform analysis, calculate the corresponding frequency offset value, and determine the offset distribution data;

[0018] acquiring initial strain distribution data based on the offset distribution data;

[0019] The initial strain distribution data is processed to obtain the strain distribution data.

[0020] Optionally, obtaining initial strain distribution data based on the offset distribution data includes:

[0021] Correcting the offset distribution data to obtain corrected offset distribution data;

[0022] Calculating the strain distribution information along the optical fiber path based on the corrected offset distribution data and combining the mapping relationship between strain distribution and frequency offset to obtain a strain distribution data set;

[0023] By performing spatial mapping on the strain distribution data set, a strain distribution map along the path is generated and the final distribution mapping result is determined;

[0024] If there are missing data or discontinuous areas in the final distribution mapping result, an interpolation method is used to supplement the missing areas to obtain the initial strain distribution data.

[0025] Optionally, processing the initial strain distribution data to obtain the strain distribution data includes:

[0026] The wavelet transform method is used to decompose the initial strain distribution data, separate the signal components of different frequencies, and obtain the decomposed multi-layer signal data;

[0027] If the high-frequency part of the decomposed multi-layer signal data exceeds a preset threshold, it is determined to be a noise interference component, and the high-frequency part is filtered out to obtain the signal data after the noise is filtered out;

[0028] The low-frequency signal portion of the strain distribution is reconstructed based on the signal data after noise is filtered out to obtain the strain distribution data.

[0029] Optionally, based on the preliminary three-dimensional deformation distribution, by acquiring real-time geological parameter data, updating the dynamic boundary conditions of the surrounding rock mechanics model includes:

[0030] updating the geological structural characteristics and material parameter inputs in the surrounding rock mechanics model using geological exploration data to generate an updated surrounding rock mechanics model;

[0031] If the stress field distribution deviation between the updated surrounding rock mechanics model and the initial model exceeds a preset threshold, the model parameters are adjusted through an iterative optimization algorithm to obtain the optimized mechanics model.

[0032] Optionally, adjusting model parameters through an iterative optimization algorithm to obtain the optimized mechanical model includes:

[0033] Obtaining geological parameters from real-time geological data, and using a preprocessing algorithm to denoise the data to obtain a first geological parameter set;

[0034] If the deviation between the first geological parameter set and the predicted value of the surrounding rock mechanics model exceeds a preset threshold, the model parameters are adjusted using a gradient descent algorithm to obtain a first optimized parameter set;

[0035] updating the surrounding rock mechanics model according to the first optimized parameter set and generating a new set of prediction values using a numerical simulation method;

[0036] If the deviation between the new predicted value set and the first geological parameter set still exceeds the preset threshold, the surrounding rock mechanics model parameters are further optimized by the Newton iteration method to obtain the second optimized parameter set;

[0037] updating the surrounding rock mechanics model according to the second optimized parameter set and generating a high-precision prediction value set using finite element analysis;

[0038] By comparing the deviation between the high-precision prediction value set and the first geological parameter set, it is determined whether a preset threshold is met, and the optimized mechanical model is obtained.

[0039] Optionally, the strain distribution data is inverted again using an optimized mechanical model to generate the final three-dimensional deformation distribution including:

[0040] For the optimized model configuration, the inverse calculation process is carried out in combination with the strain distribution data. The finite element analysis method is used to handle the three-dimensional spatial analysis task and perform error analysis to obtain the corrected deformation distribution.

[0041] The final distribution result is generated through the corrected deformation distribution, and the deformation distribution diagram is visualized using three-dimensional spatial analysis technology to obtain the final three-dimensional deformation distribution view;

[0042] If the final three-dimensional deformation distribution view has missing data, supplementary data is obtained from the strain data source, and a local update is performed in combination with the inversion calculation process to generate the final three-dimensional deformation distribution.

[0043] Optionally, based on the final three-dimensional deformation distribution, a comparison is performed using a preset deformation threshold to determine the stability of the surrounding rock mass and generate an early warning signal, including:

[0044] The final three-dimensional deformation distribution is compared point by point with a preset deformation threshold to obtain a preliminary abnormal deformation area distribution;

[0045] According to the distribution of the preliminary abnormal deformation areas, the abnormal areas are grouped using a spatial clustering method to determine the target area range of concentrated deformation;

[0046] Obtaining time series variation characteristics of the three-dimensional deformation data within the target area to determine whether the deformation trend continues to intensify;

[0047] Based on the judgment results, an early warning signal is generated.

[0048] Optionally, generating a warning signal according to the judgment result includes:

[0049] If the deformation trend continues to intensify, the stability index is calculated in combination with the surrounding rock stability analysis model to obtain the potential risk level of the surrounding rock;

[0050] According to the potential risk level, a corresponding warning signal level is generated to determine whether to trigger an emergency response mechanism;

[0051] The final warning information output content is obtained by matching the warning signal level with the preset signal generation rules.

[0052] Compared with the prior art, the present invention has the following advantages and technical effects:

[0053] The present invention obtains backscattered light signals by placing fiber optic sensors on the surface of the tunnel surrounding rock, and uses the Brillouin optical time domain analysis algorithm to extract strain distribution data. In response to the interference of complex geological environments, wavelet transform denoising is used. Combined with the pre-established surrounding rock mechanical model, the finite element method is used for inversion calculation to obtain the three-dimensional deformation distribution. The present invention also updates the mechanical model through real-time geological parameters and uses an iterative optimization algorithm to improve the model accuracy. Finally, based on the comparison of the deformation distribution with the preset threshold, the stability state of the surrounding rock is judged and an early warning signal is generated. This method realizes real-time and high-precision monitoring of tunnel surrounding rock deformation, providing a strong guarantee for mine safety production. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0055] Figure 1 This is a flow chart of a method for monitoring deformation of surrounding rock in a tunnel based on an optical fiber sensor according to an embodiment of the present invention. DETAILED DESCRIPTION

[0056] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0057] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0058] This embodiment proposes a tunnel surrounding rock deformation monitoring method based on optical fiber sensors. Figure 1 As shown, the specific steps include:

[0059] Distributed optical fiber sensing technology is deployed on the surface of the tunnel surrounding rock to obtain raw optical signal data;

[0060] The backscattered light data in the original optical signal data is processed by the Brillouin optical time domain analysis algorithm to extract the frequency offset of the optical signal and obtain the strain distribution data;

[0061] Based on the strain distribution data and the pre-established surrounding rock mechanical model, the finite element method is used to perform initial inversion calculations to obtain the preliminary three-dimensional deformation distribution;

[0062] Based on the preliminary three-dimensional deformation distribution, the dynamic boundary conditions of the surrounding rock mechanical model are updated by obtaining real-time geological parameter data to obtain an optimized mechanical model;

[0063] The strain distribution data is inverted again using the optimized mechanical model to generate the final three-dimensional deformation distribution;

[0064] Based on the final three-dimensional deformation distribution, a comparison and judgment is made through the preset deformation threshold to determine the stability status of the surrounding rock and generate an early warning signal.

[0065] Specifically, the present invention obtains backscattered light signals by placing fiber optic sensors on the surface of the tunnel surrounding rock, and uses the Brillouin optical time domain analysis algorithm to extract strain distribution data. In response to the interference of complex geological environments, wavelet transform denoising is used. Combined with the pre-established surrounding rock mechanical model, the finite element method is used for inversion calculation to obtain the three-dimensional deformation distribution. The present invention also updates the mechanical model through real-time geological parameters and uses an iterative optimization algorithm to improve the model accuracy. Finally, based on the comparison of the deformation distribution with the preset threshold, the stability state of the surrounding rock is judged and an early warning signal is generated. This method realizes real-time and high-precision monitoring of tunnel surrounding rock deformation, providing a strong guarantee for mine safety production.

[0066] Furthermore, distributed optical fiber sensing technology is deployed on the surface of the tunnel surrounding rock to obtain raw optical signal data including:

[0067] A sensor network is deployed on the surface of the tunnel surrounding rock using distributed optical fiber sensing technology to obtain initial optical signal data containing continuous backscattered light signals.

[0068] The wavelet transform algorithm is used to perform denoising on the initial optical signal data to obtain the original optical signal data.

[0069] Specifically, when using distributed fiber optic sensing technology to monitor the surrounding rock conditions of a tunnel, a fiber optic sensor network can be deployed on the tunnel surface to collect backscattered light signals in real time. Consider a mine tunnel with a 500-meter-long sensing fiber, with monitoring points set every meter. The raw optical signal data obtained may contain abnormal fluctuations caused by environmental noise or equipment interference. To address this noise issue, a wavelet transform algorithm is used for denoising. This decomposes the signal into frequency components, filters out high-frequency noise, and retains the low-frequency, valid signal, resulting in denoised optical signal data. This method effectively improves the signal-to-noise ratio and provides a more reliable data foundation for subsequent analysis. For example, during the feature parameter extraction stage, the signal amplitude in the time domain and the dominant frequency component in the frequency domain can be extracted from the denoised optical signal data. Suppose that the signal amplitude at a certain fiber optic monitoring point suddenly increases to twice its normal value, and the dominant frequency shifts from 0.5 Hz to 1.2 Hz, exceeding the preset threshold. At this point, analysis of the spectrum using the Fourier transform algorithm revealed that the abnormal signal points were concentrated approximately 200 meters in the middle of the tunnel. The spectrum displayed a distinct abnormal peak, indicating possible localized stress concentration in the surrounding rock or the expansion of microcracks. The application of the Fourier transform helps precisely locate the frequency distribution of abnormalities, improving the accuracy of anomaly detection. For example, in the classification of abnormal signals, the K-means clustering algorithm was used to classify the abnormal signal points. Assume that the detected anomalies fall into two categories: one with high amplitude but stable frequency, and the other with wildly fluctuating frequency. Cluster analysis can classify these two categories as surrounding rock stress accumulation and dynamic deformation, respectively. This classification method helps distinguish different types of surrounding rock anomalies, providing a basis for subsequent spatial location and improving the targeted nature of monitoring. For example, when determining the spatial distribution of abnormal signals, combined with the spatial distribution data of the sensor network, it is assumed that the abnormal signals are concentrated between 200 and 250 meters in the middle of the tunnel and exhibit a linear distribution pattern, suggesting that they may be related to the geological structure or mining activities in this area. By generating real-time monitoring data, a trend chart of surrounding rock state changes can be drawn, and it was found that the stress value in this area has continued to rise in the past 24 hours, indicating a potential deformation risk. This spatial positioning and trend analysis can provide an intuitive basis for tunnel safety management, allowing timely reinforcement measures to reduce accident risks. For example, in terms of overall monitoring effects, through the comprehensive application of the above-mentioned technical means, not only can continuous monitoring of surrounding rock conditions be achieved, but also abnormal areas and types can be accurately identified, significantly improving monitoring efficiency and early warning capabilities. Especially under complex geological conditions, this method can effectively deal with signal interference and data redundancy problems, ensure the reliability of monitoring data, and provide technical support for safe mine operations.

[0070] Furthermore, the backscattered light data in the original optical signal data is processed by the Brillouin optical time domain analysis algorithm to extract the frequency offset of the optical signal and obtain the strain distribution data, including:

[0071] By collecting and preprocessing the optical signal data, the backscattered signal is separated from the original data to obtain the initial scattered signal data;

[0072] The Brillouin optical time domain analysis method is used to perform time domain processing on the initial scattered signal data, extract characteristic information related to the frequency offset, and obtain the frequency offset feature set.

[0073] Based on the frequency offset feature set, a preset offset calculation model is used to perform analysis, calculate the corresponding frequency offset value, and determine the offset distribution data;

[0074] Based on the offset distribution data, initial strain distribution data is obtained;

[0075] The initial strain distribution data is processed to obtain strain distribution data.

[0076] Specifically, during the optical signal data acquisition and preprocessing process, the raw data can be initially filtered to isolate the backscattered signal. For example, in a tunnel surrounding rock monitoring scenario, where fiber optic sensors are deployed along the tunnel surface, the collected raw data may contain environmental noise and equipment interference. In this case, time-domain sampling can be used to segment the signal into time-sequential segments, extracting the signal portion related to backscattering and generating initial scattered signal data. This approach effectively focuses on the target signal and lays the foundation for subsequent analysis. For example, the application of Brillouin optical time-domain analysis can be understood as an analysis method based on the interaction between light and sound waves in optical fibers. In tunnel monitoring, this method extracts characteristic information related to frequency offset by performing time-domain processing on the initial scattered signal data. Assuming that the optical fiber is deployed for 1000 meters along the tunnel, analyzing the signal time-domain characteristics of each fiber segment can yield a frequency offset feature set, reflecting the stress conditions at different locations on the fiber. This feature extraction facilitates the subsequent precise location of abnormal areas. For example, when analyzing the frequency offset feature set using a pre-set offset calculation model, the specific frequency offset value can be calculated using the model trained using historical data. Assuming the frequency offset of a certain fiber section is 50 Hz, combined with model analysis, we can determine the offset distribution data. This distribution data can intuitively reflect the variation trend along the fiber and provide a basis for subsequent strain calculations.

[0077] Furthermore, based on the offset distribution data, obtaining initial strain distribution data includes:

[0078] Correcting the offset distribution data to obtain corrected offset distribution data;

[0079] According to the corrected offset distribution data, combined with the mapping relationship between strain distribution and frequency offset, the strain distribution information along the optical fiber path is calculated to obtain a strain distribution data set;

[0080] By performing spatial mapping on the strain distribution data set, a strain distribution map along the path is generated and the final distribution mapping result is determined;

[0081] If there are missing data or discontinuous areas in the final distribution mapping results, the interpolation method is used to supplement the missing areas to obtain the initial strain distribution data.

[0082] Specifically, if the offset distribution data contains anomalies, such as a point where the offset value suddenly reaches 200 Hz, far exceeding the normal range of 10-60 Hz, it is filtered using a preset threshold range to remove the outliers and obtain corrected offset distribution data. This filtering method can improve data reliability and avoid misjudgments. For example, when calculating strain distribution information from the corrected offset distribution data, the strain distribution along the optical fiber path can be calculated based on the mapping relationship between frequency offset and strain, assuming that every 10 Hz increase in frequency offset corresponds to a 20 microstrain increase in strain. This strain distribution dataset provides data support for subsequent spatial mapping. For example, by performing spatial mapping on the strain distribution dataset, a strain distribution map along the path can be generated. Assuming that the strain value at a certain 100-meter section of the tunnel is high, reaching 100 microstrain, the distribution map can intuitively demonstrate the strain concentration in this area. This visualization facilitates the rapid identification of potential risk points. For example, if the final distribution mapping result contains missing data, such as incomplete acquisition of a certain section of optical fiber signal, interpolation can be used to supplement the missing areas. Assuming data is missing between 200 and 210 meters, linear interpolation can be performed based on the preceding and following data trends to obtain complete strain distribution mapping data. This method ensures data continuity and enhances comprehensive monitoring. The above multifaceted processing and analysis demonstrate the important role of each technical topic in roadway surrounding rock monitoring. These methods support each other, ensuring the integrity and accuracy of the entire process from signal acquisition to final mapping, providing a reliable basis for real-time monitoring.

[0083] Furthermore, the initial strain distribution data is processed to obtain the strain distribution data, including:

[0084] The wavelet transform method is used to decompose the initial strain distribution data, separate the signal components of different frequencies, and obtain the decomposed multi-layer signal data;

[0085] If the high-frequency part of the decomposed multi-layer signal data exceeds a preset threshold, it is determined to be a noise interference component, and the high-frequency part is filtered out to obtain the signal data after the noise is filtered out;

[0086] Based on the signal data after noise filtering, the low-frequency signal part of the strain distribution is reconstructed to obtain the strain distribution data.

[0087] Specifically, when processing raw strain distribution data, the initially collected data often contains various noise artifacts, such as environmental vibrations or subtle errors in the equipment itself. To address this issue, the initial dataset can be decomposed using wavelet transforms. The principle of wavelet transforms is to decompose a signal into components of different frequencies to facilitate the distinction between valid signals and noise. For example, in one possible implementation, assume that the collected strain data contains signals with frequencies ranging from 0.1 Hz to 50 Hz. Wavelet decomposition can separate the signals into multiple layers, where the high-frequency components typically correspond to noise, while the low-frequency components reflect actual strain changes. After decomposition, if the amplitude of a high-frequency signal in a layer exceeds a preset threshold, such as 0.05, it can be determined to be noise interference and filtered out, retaining the valid low-frequency data. For example, when reconstructing the strain distribution from the noise-filtered signal data, the reconstruction of the low-frequency signals can be focused on. Low-frequency signals typically represent relatively smooth strain variations along the optical fiber, while high-frequency noise may originate from transient interference. For example, for a given fiber path, the reconstructed strain values fluctuate between 0.001 and 0.003. This data better reflects the strain characteristics of the actual geological environment. Reconstruction effectively restores the true strain distribution, providing a reliable foundation for subsequent analysis.

[0088] Furthermore, based on the preliminary 3D deformation distribution and by acquiring real-time geological parameter data, the dynamic boundary conditions of the surrounding rock mechanics model are updated, including:

[0089] Using geological exploration data, the geological structural characteristics and material parameter inputs in the surrounding rock mechanics model are updated to generate an updated surrounding rock mechanics model;

[0090] If the stress field distribution deviation between the updated surrounding rock mechanical model and the initial model exceeds a preset threshold, the model parameters are adjusted through an iterative optimization algorithm to obtain an optimized mechanical model.

[0091] Specifically, when using the finite element method for rock mechanics analysis, a pre-built model can be combined with existing strain distribution data to simulate the deformation behavior of the surrounding rock under specific conditions. The core of the finite element method is to discretize the complex surrounding rock structure into multiple small units. By calculating the stress and deformation of each unit, the overall three-dimensional deformation distribution is gradually derived. Suppose, for example, in a tunnel project, the initial model assumes the surrounding rock is homogeneous. After calculating the preliminary deformation distribution, it is found that some areas have excessive deformation, which may not be consistent with the actual geological conditions. For example, to update geological exploration data, the actual lithology parameters and fracture distribution of the surrounding rock can be obtained through on-site drilling. Suppose the exploration data indicates the presence of weak interlayers in the surrounding rock. The material strength parameter needs to be adjusted from the initial 50 MPa to 30 MPa. At the same time, the geological structural characteristics are updated, and the uniform distribution in the model is adjusted to a layered distribution. This updated rock mechanics model is more realistic and can improve the reliability of subsequent calculations. For example, when the stress field distribution deviation exceeds a preset threshold, it is particularly important to adjust the boundary conditions.

[0092] Furthermore, the model parameters are adjusted through iterative optimization algorithm to obtain the optimized mechanical model including:

[0093] Obtaining geological parameters from real-time geological data, and using a preprocessing algorithm to denoise the data to obtain a first geological parameter set;

[0094] If the deviation between the first geological parameter set and the predicted value of the surrounding rock mechanics model exceeds a preset threshold, the model parameters are adjusted using a gradient descent algorithm to obtain a first optimized parameter set;

[0095] updating the surrounding rock mechanics model according to the first optimized parameter set and generating a new set of prediction values using a numerical simulation method;

[0096] If the deviation between the new predicted value set and the first geological parameter set still exceeds the preset threshold, the surrounding rock mechanics model parameters are further optimized by the Newton iteration method to obtain the second optimized parameter set;

[0097] updating the surrounding rock mechanics model according to the second optimized parameter set and generating a high-precision prediction value set using finite element analysis;

[0098] By comparing the deviation between the high-precision prediction value set and the first geological parameter set, it is determined whether the preset threshold is met and the optimized mechanical model is obtained.

[0099] Specifically, when acquiring real-time geological data to construct the first geological parameter set, data such as pressure, temperature, and porosity of the underground rock formation can be collected through a sensor network. Suppose an oilfield monitoring system collects 100 sets of pressure data per minute. The raw data may contain outliers due to equipment noise or environmental interference. A preprocessing algorithm can use a median filter to remove outliers outside the normal range. For example, data with pressure values exceeding 200 MPa is considered noise. This processing yields the first geological parameter set, which contains stable and reliable pressure distribution data. This denoising method effectively improves data quality and ensures the accuracy of subsequent analysis. In one possible implementation, if the deviation between the first geological parameter set and the mechanical model prediction exceeds a preset threshold (e.g., 5%), the model parameters need to be adjusted. A gradient descent algorithm can be used to optimize the elastic modulus and Poisson's ratio in the model. For example, the initial model assumes a rock formation elastic modulus of 30 GPa, but the actual geological parameters show 32 GPa. Through multiple iterations, the algorithm gradually adjusts the values closer to the actual values, generating the first optimized parameter set. This method can quickly converge to a reasonable parameter range, enhancing the model's adaptability to complex geological environments. Specifically, after updating the mechanical model based on the first optimized parameter set, numerical simulation can be used to generate a new set of predicted values. Suppose a simulation predicts a stress distribution of 10 MPa in a certain area, while the actual monitored value is 12 MPa. The deviation still exceeds the threshold. Further optimization is required. The Newton iteration method can be used to fine-tune model parameters, such as adjusting the Poisson's ratio from 0.25 to 0.27 through iterative calculations. Compared to gradient descent, the Newton iteration method is more efficient when approaching the optimal solution, significantly improving model accuracy. For example, after updating the mechanical model with the second optimized parameter set, finite element analysis can be used to generate a high-precision set of predicted values. Finite element analysis divides the geological body into 100,000 grid cells and calculates the stress and deformation of each cell. Suppose the predicted value indicates a maximum deformation of 2 mm in a certain area, while the actual monitored value is 2.1 mm. The deviation is reduced to within 1%, meeting the preset threshold. This high-precision prediction better reflects the true state of the geological body and provides a reliable basis for subsequent engineering decisions. In one possible implementation, the deviation between the high-precision predicted value set and the first geological parameter set is compared to determine whether it meets the preset threshold, thereby determining the final mechanical model. For example, deviation analysis showed that 90% of the predicted values differed from the actual values by less than 2%, demonstrating that the model closely matches the actual geological conditions. The final model can be used to analyze formation stability and optimize engineering designs. This deviation analysis method effectively verifies the reliability of the model and ensures its practicality in real-world applications.

[0100] Furthermore, the strain distribution data is inverted again using the optimized mechanical model to generate the final three-dimensional deformation distribution including:

[0101] For the optimized model configuration, the inverse calculation process is carried out in combination with the strain distribution data. The finite element analysis method is used to handle the three-dimensional spatial analysis task and perform error analysis to obtain the corrected deformation distribution.

[0102] The final distribution result is generated through the corrected deformation distribution, and the deformation distribution diagram is visualized using three-dimensional spatial analysis technology to obtain the final three-dimensional deformation distribution view;

[0103] If there is data missing in the final three-dimensional deformation distribution view, supplementary data is obtained from the strain data source and locally updated in combination with the inversion calculation process to generate the final three-dimensional deformation distribution.

[0104] Furthermore, based on the final three-dimensional deformation distribution, a comparison is made through a preset deformation threshold to determine the stability of the surrounding rock mass and generate an early warning signal, including:

[0105] The final three-dimensional deformation distribution is compared point by point with the preset deformation threshold to obtain the preliminary deformation abnormal area distribution;

[0106] According to the preliminary distribution of abnormal deformation areas, the spatial clustering method is used to group the abnormal areas and determine the target area of concentrated deformation;

[0107] Based on the three-dimensional deformation data within the target area, obtain its time series change characteristics and determine whether the deformation trend continues to intensify;

[0108] Based on the judgment results, an early warning signal is generated.

[0109] Furthermore, generating an early warning signal according to the judgment result includes:

[0110] If the deformation trend continues to intensify, the stability index is calculated in combination with the surrounding rock stability analysis model to obtain the potential risk level of the surrounding rock;

[0111] Generate corresponding warning signal levels based on potential risk levels to determine whether to trigger emergency response mechanisms;

[0112] By matching the warning signal level with the preset signal generation rules, the final warning information output content is obtained.

[0113] Specifically, when analyzing 3D deformation data, one can begin with point-by-point comparison. For example, in an underground engineering monitoring project, the acquired 3D deformation data contains displacement values from 1,000 monitoring points. These values are compared against a preset deformation threshold of 0.5 cm. Fifty points are found to exceed the threshold, and the areas located in these points are preliminarily identified as areas of abnormal deformation. This point-by-point comparison method can quickly identify potential problem areas, laying the foundation for subsequent analysis. For example, using spatial clustering to identify the initial distribution of abnormal deformation areas, these 50 abnormal points can be grouped by spatial distance. For example, clustering with a radius of 5 meters will ultimately identify three key areas of concentrated deformation. This method effectively reduces the interference of scattered abnormal points, focusing on areas potentially posing greater risks and improving the targeted nature of the analysis. For example, when extracting time series characteristics of key areas, seven consecutive days of monitoring data can be used to observe whether the deformation increases daily. Suppose the average deformation in a certain area increases from 0.6 cm to 0.9 cm, indicating a continuously increasing trend. This trend analysis helps determine the severity of a problem and provides data support for subsequent risk assessments. For example, when combined with a surrounding rock stability analysis model, a stability index can be calculated based on time series data. For example, if the stability index of a key area is 0.7, below the safe value of 1.0, indicating a certain risk. This index-based analysis quantifies the surrounding rock condition and facilitates tiered management. For example, when generating warning signal levels, if the stability index is below 0.8, it can be classified as medium risk, triggering a yellow warning signal. This grading mechanism intuitively reflects the degree of risk and facilitates rapid response by relevant personnel. For example, the output content of warning information can be formatted according to preset rules, and medium-risk warning information can be formatted into a brief summary that includes the time, location, and risk description. This standardized output method ensures clear and consistent information transmission. For example, if the warning information indicates a high risk level, the system can automatically record the deformation data of the area, such as a maximum deformation of 1.2 cm and a stability index of 0.5, and generate a monitoring log. This automatic recording mechanism provides a comprehensive basis for subsequent review and decision-making, helping to improve the reliability of the monitoring system.

[0114] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for monitoring deformation of surrounding rock in tunnels based on optical fiber sensors, characterized in that: include: Distributed optical fiber sensing technology is deployed on the surface of the tunnel surrounding rock to obtain raw optical signal data; Processing the backscattered light data in the original optical signal data by using a Brillouin optical time domain analysis algorithm, extracting the frequency offset of the optical signal, and obtaining strain distribution data; Based on the strain distribution data and the pre-established surrounding rock mechanical model, the finite element method is used to perform initial inversion calculations to obtain the preliminary three-dimensional deformation distribution; Based on the preliminary three-dimensional deformation distribution, the dynamic boundary conditions of the surrounding rock mechanical model are updated by acquiring real-time geological parameter data to obtain an optimized mechanical model; The strain distribution data is inverted again using the optimized mechanical model to generate the final three-dimensional deformation distribution; Based on the final three-dimensional deformation distribution, a comparison and judgment is made through the preset deformation threshold to determine the stability status of the surrounding rock and generate an early warning signal.

2. The method for monitoring deformation of surrounding rock in tunnels based on optical fiber sensors according to claim 1, characterized in that: Distributed optical fiber sensing technology is deployed on the surface of the tunnel surrounding rock to obtain raw optical signal data including: A sensor network is deployed on the surface of the tunnel surrounding rock using distributed optical fiber sensing technology to obtain initial optical signal data containing continuous backscattered light signals. The wavelet transform algorithm is used to perform denoising on the initial optical signal data to obtain the original optical signal data.

3. The method for monitoring deformation of surrounding rock in tunnels based on optical fiber sensors according to claim 1, characterized in that: Processing the backscattered light data in the original optical signal data by using a Brillouin optical time domain analysis algorithm, extracting the frequency offset of the optical signal, and obtaining the strain distribution data includes: By collecting and preprocessing the optical signal data, the backscattered signal is separated from the original data to obtain the initial scattered signal data; The Brillouin optical time domain analysis method is used to perform time domain processing on the initial scattered signal data, extract characteristic information related to the frequency offset, and obtain the frequency offset feature set. Based on the frequency offset feature set, a preset offset calculation model is used to perform analysis, calculate the corresponding frequency offset value, and determine the offset distribution data; acquiring initial strain distribution data based on the offset distribution data; The initial strain distribution data is processed to obtain the strain distribution data.

4. The method for monitoring deformation of surrounding rock in tunnels based on optical fiber sensors according to claim 3, characterized in that: Acquiring initial strain distribution data based on the offset distribution data includes: Correcting the offset distribution data to obtain corrected offset distribution data; Calculating the strain distribution information along the optical fiber path based on the corrected offset distribution data and combining the mapping relationship between strain distribution and frequency offset to obtain a strain distribution data set; By performing spatial mapping on the strain distribution data set, a strain distribution map along the path is generated and the final distribution mapping result is determined; If there are missing data or discontinuous areas in the final distribution mapping result, an interpolation method is used to supplement the missing areas to obtain the initial strain distribution data.

5. The method for monitoring deformation of surrounding rock in tunnels based on optical fiber sensors according to claim 3, characterized in that: Processing the initial strain distribution data to obtain the strain distribution data includes: The wavelet transform method is used to decompose the initial strain distribution data, separate the signal components of different frequencies, and obtain the decomposed multi-layer signal data; If the high-frequency part of the decomposed multi-layer signal data exceeds a preset threshold, it is determined to be a noise interference component, and the high-frequency part is filtered out to obtain the signal data after the noise is filtered out; The low-frequency signal portion of the strain distribution is reconstructed based on the signal data after noise is filtered out to obtain the strain distribution data.

6. The method for monitoring deformation of surrounding rock in tunnels based on optical fiber sensors according to claim 1, characterized in that: Based on the preliminary three-dimensional deformation distribution, by acquiring real-time geological parameter data, the dynamic boundary conditions of the surrounding rock mechanics model are updated, including: updating the geological structural characteristics and material parameter inputs in the surrounding rock mechanics model using geological exploration data to generate an updated surrounding rock mechanics model; If the stress field distribution deviation between the updated surrounding rock mechanics model and the initial model exceeds a preset threshold, the model parameters are adjusted through an iterative optimization algorithm to obtain the optimized mechanics model.

7. The method for monitoring deformation of surrounding rock in tunnels based on optical fiber sensors according to claim 6, characterized in that: The optimized mechanical model is obtained by adjusting the model parameters through an iterative optimization algorithm, including: Obtaining geological parameters from real-time geological data, and using a preprocessing algorithm to denoise the data to obtain a first geological parameter set; If the deviation between the first geological parameter set and the predicted value of the surrounding rock mechanics model exceeds a preset threshold, the model parameters are adjusted using a gradient descent algorithm to obtain a first optimized parameter set; updating the surrounding rock mechanics model according to the first optimized parameter set and generating a new set of prediction values using a numerical simulation method; If the deviation between the new predicted value set and the first geological parameter set still exceeds the preset threshold, the surrounding rock mechanics model parameters are further optimized by the Newton iteration method to obtain the second optimized parameter set; updating the surrounding rock mechanics model according to the second optimized parameter set and generating a high-precision prediction value set using finite element analysis; By comparing the deviation between the high-precision prediction value set and the first geological parameter set, it is determined whether a preset threshold is met, and the optimized mechanical model is obtained.

8. The method for monitoring deformation of surrounding rock in tunnels based on optical fiber sensors according to claim 1, characterized in that: The strain distribution data is inverted again using the optimized mechanical model to generate the final three-dimensional deformation distribution including: For the optimized model configuration, the inverse calculation process is carried out in combination with the strain distribution data. The finite element analysis method is used to handle the three-dimensional spatial analysis task and perform error analysis to obtain the corrected deformation distribution. The final distribution result is generated through the corrected deformation distribution, and the deformation distribution diagram is visualized using three-dimensional spatial analysis technology to obtain the final three-dimensional deformation distribution view; If the final three-dimensional deformation distribution view has missing data, supplementary data is obtained from the strain data source, and a local update is performed in combination with the inversion calculation process to generate the final three-dimensional deformation distribution.

9. The method for monitoring deformation of surrounding rock in tunnels based on optical fiber sensors according to claim 1, characterized in that: Based on the final 3D deformation distribution, the preset deformation threshold is used for comparison and judgment to determine the stability of the surrounding rock mass and generate early warning signals, including: The final three-dimensional deformation distribution is compared point by point with a preset deformation threshold to obtain a preliminary abnormal deformation area distribution; According to the distribution of the preliminary abnormal deformation areas, the abnormal areas are grouped using a spatial clustering method to determine the target area range of concentrated deformation; Obtaining time series variation characteristics of the three-dimensional deformation data within the target area to determine whether the deformation trend continues to intensify; Based on the judgment results, an early warning signal is generated.

10. The method for monitoring deformation of surrounding rock in tunnels based on optical fiber sensors according to claim 9, characterized in that: Generating an early warning signal according to the judgment result includes: If the deformation trend continues to intensify, the stability index is calculated in combination with the surrounding rock stability analysis model to obtain the potential risk level of the surrounding rock; According to the potential risk level, a corresponding warning signal level is generated to determine whether to trigger an emergency response mechanism; The final warning information output content is obtained by matching the warning signal level with the preset signal generation rules.

Images