A method to improve the accuracy of surrounding rock deformation detection

Abstract

This application relates to the field of surrounding rock detection technology, specifically to a method for improving the accuracy of surrounding rock deformation detection. The method includes: acquiring surrounding rock monitoring data; calculating a geological difference coefficient based on the deformation differences between different monitoring points to measure the differences in geological uniformity between different locations of the surrounding rock; calculating an environmental correlation coefficient using the correlation between the deformation at each monitoring point and environmental changes to measure the impact of environmental changes on the deformation at different locations of the surrounding rock; determining a disturbance coefficient at each monitoring point using the correspondence between the geological difference coefficient and the environmental correlation coefficient to measure the degree of inhomogeneity of the surrounding rock area and the impact of environmental changes on the deformation of the surrounding rock; and using the disturbance coefficient to predict strain data and correct the strain data acquired by the sensors to complete the detection of surrounding rock deformation. This application aims to improve the accuracy of surrounding rock deformation detection results.

Description

Technical Field

[0001] This application relates to the field of surrounding rock detection technology, specifically to a method for improving the accuracy of surrounding rock deformation detection. Background Technology

[0002] With the increasing development and utilization of underground space, such as the growing number of tunnel projects in subways, highways, and mines, the stability of the surrounding rock, as a crucial component of underground structures, is vital to the overall safety of the project. Therefore, monitoring surrounding rock deformation becomes particularly important in these projects.

[0003] Current technologies for detecting surrounding rock deformation primarily rely on the differences between detected data and historical or normal data to determine whether deformation has occurred. This method assumes that the surrounding rock area is uniformly distributed, but in reality, surrounding rock areas often exhibit non-uniform distribution characteristics. Under the influence of different geological conditions and environmental changes, the impact of different parts of the surrounding rock on the sensor varies, and the rate and intensity of change also differ significantly. This makes it difficult for the sensor to accurately assess the actual situation at each location when using existing methods, resulting in errors in the detection results. Summary of the Invention

[0004] To address the aforementioned technical problems, this application provides a method for improving the accuracy of surrounding rock deformation detection, thereby resolving the existing issues.

[0005] The method for improving the accuracy of surrounding rock deformation detection in this application adopts the following technical solution:

[0006] One embodiment of this application provides a method for improving the accuracy of surrounding rock deformation detection, the method comprising the following steps:

[0007] Pre-set monitoring points at each location of the surrounding rock, perform curve fitting on the strain data obtained from each monitoring point in each test and historical tests, analyze the shape characteristics of the fitted curves of all monitoring points in each test, and the fluctuation of strain data of all monitoring points at the same location in each test, and determine the geological difference coefficient of each test.

[0008] The ambient temperature of each test and historical tests is mapped onto a coordinate system, and the lines are connected sequentially to obtain a temperature curve. The area corresponding to the temperature curve above the horizontal axis is taken as the freeze-thaw region. Based on the similarity between the strain data of each monitoring point in the non-freeze-thaw region and the ambient humidity in the corresponding freeze-thaw region, the similarity between the ambient humidity of each test and historical tests and the strain data of each monitoring point, combined with the shape characteristics of the freeze-thaw region and the distribution differences of the strain data of the monitoring points in each test, the environmental correlation coefficient of each monitoring point in each test is determined.

[0009] For each test, based on the geological difference coefficient and the environmental correlation coefficient of all monitoring points, the disturbance coefficient of each monitoring point is determined. Combined with the regression model, the strain prediction value for the next test is obtained. The strain data is then corrected to obtain the test results of the surrounding rock deformation.

[0010] The determination of the geological difference coefficient for each test includes:

[0011] Analyze the shape differences of the fitted curves at each monitoring point during different detections to determine the strain trend at each monitoring point for each detection.

[0012] Analyze the fluctuation of strain data at all monitoring points at the same location for each test to obtain a measure of the anchor bolt strain trend for each test;

[0013] Obtain the cumulative sum of strain trend values ​​for all monitoring points in each test; calculate the mean of the determination coefficients of the fitted curves for all monitoring points in each test, and add it to the anchor strain trend metric; perform negative correlation mapping on the summed values ​​and positively fuse them with the cumulative sum to obtain the geological difference coefficient for each test.

[0014] The determination of the strain trend at each monitoring point for each detection includes:

[0015] First, for each detection, obtain the integral value formed by the fitted curve function and the coordinate system for each monitoring point;

[0016] Obtain the integral value sequence of each monitoring point, which consists of the integral values ​​of each detection and all previous detections; obtain the first-order difference sequence of the integral value sequence, substitute the negative numbers of all elements in the first-order difference sequence into the sign function and average them, and then perform positive fusion with the integral value of each detection to obtain the strain trend quantity of each monitoring point in each detection.

[0017] The method for obtaining the anchor strain trend measurement for each test includes:

[0018] The strain data of all monitoring points at the same location at each test is recorded as the strain sequence for each location;

[0019] For each detection, the monitoring point corresponding to the minimum value of the strain sequence at each location is obtained as a reference point;

[0020] The percentage of negative differences in strain data between all adjacent monitoring points preceding the reference point is recorded as the first percentage.

[0021] The percentage of positive differences in strain data between all adjacent monitoring points following the reference point is recorded as the second percentage. The average of the sum of the first percentage and the second percentage for all locations in each test and all previous tests is used as the anchor strain trend measure for each test.

[0022] The determination of the environmental correlation coefficient for each monitoring point in each detection includes:

[0023] Based on the distribution of strain data at each monitoring point in the corresponding section of the non-freeze-thaw zone, the freeze-thaw strain sequence of each monitoring point is obtained;

[0024] The distribution characteristics of ambient humidity in the corresponding sections of the freeze-thaw region were analyzed, and the freeze-thaw sequence was obtained by combining the area of ​​the freeze-thaw region.

[0025] For each monitoring point, calculate the difference in strain data between each monitoring point and all other monitoring points in each test, and calculate the average level of the difference in strain data between each monitoring point and all other monitoring points in each test; use the ratio of the average level of each monitoring point in each test to the average level of the next test as the difference coefficient of each monitoring point in each test.

[0026] The correlation measure between the humidity sequence and the strain time series obtained in each detection for each monitoring point is obtained and denoted as the first correlation. The correlation measure between the freeze-thaw strain sequence and the freeze-thaw sequence obtained in each detection for each monitoring point is obtained and denoted as the second correlation. The sum of the first correlation and the second correlation for each monitoring point in each detection is calculated. The negative correlation mapping result of the average difference coefficient obtained by each monitoring point in each detection and all previous detections is calculated and positively fused with the sum to obtain the environmental correlation coefficient of each monitoring point in each detection.

[0027] The process of obtaining the freeze-thaw strain sequence for each monitoring point includes:

[0028] Calculate the average strain of all strain data at each monitoring point within the corresponding segment of each non-freeze-thaw region, and combine the average strain of each monitoring point across all non-freeze-thaw regions to form the freeze-thaw strain sequence for each monitoring point.

[0029] The step of obtaining the freeze-thaw sequence is as follows:

[0030] Calculate the product between the area of ​​each freeze-thaw region and the average value of the ambient humidity in its corresponding horizontal axis interval. Use this product as the freeze-thaw coefficient of the corresponding freeze-thaw region. Combine the freeze-thaw coefficients obtained from each detection and all previous detections to form a freeze-thaw sequence.

[0031] Specifically, determining the disturbance coefficient for each monitoring point in each detection involves:

[0032] The formula for the disturbance coefficient is: In the formula, , They are the first sequence Monitoring points during the second test The disturbance coefficient, , These are the geological difference coefficient of the i-th detection and the environmental correlation coefficient of the m-th monitoring point in the i-th detection, respectively. It was before The maximum value of the geological difference coefficient in this test. It is the first The maximum value among all geological difference coefficients at all monitoring points during the second test.

[0033] The specific process of obtaining the predicted strain value for the next detection by combining the regression model is as follows:

[0034] The ratio of the disturbance coefficient of each monitoring point in each detection to the maximum value of the disturbance coefficient of all monitoring points in each detection is used as the normalized result of the disturbance coefficient of each monitoring point in each detection; the normalized result is used as the weight of the strain data of each monitoring point in each detection, and the weights of the ambient temperature and ambient humidity in each detection are preset.

[0035] A regression model is trained using a preset number of samples. Based on the data obtained from each monitoring point in each detection and historical detections, the strain prediction value for the next detection is obtained.

[0036] The step of correcting the strain data based on the error between the strain data detected each time and the predicted strain value to obtain the detection result of the surrounding rock deformation includes:

[0037] Each monitoring point will be on the [number]th The corrected strain data from the next test is denoted as Its formula is as follows: In the formula: Is each monitoring point on the 1st The strain data corrected during the second test. These are the strain data of each monitoring point during the i-th measurement. It is the preset attenuation factor. It is the first Error during the second test , They are the first sequence Correction error for the second test;

[0038] A threshold segmentation algorithm is used to obtain a segmentation threshold from the strain data of a preset number of samples. When the corrected strain data is greater than the segmentation threshold, it is determined that there is no surrounding rock deformation at the corresponding monitoring point; otherwise, there is no surrounding rock deformation.

[0039] This application has at least the following beneficial effects:

[0040] This application first calculates the geological difference coefficient based on the strain conditions at different monitoring points in the surrounding rock area. This index analyzes the strain differences at different locations caused by variations in geological structure, which helps to quantify the inhomogeneity of the geological structure and thus improves the accuracy of subsequent disturbance analysis. Then, it calculates the environmental correlation coefficient by utilizing the relationship between the deformation of the monitoring points and environmental changes. This coefficient measures the degree of environmental influence on different monitoring points and also helps to further improve the subsequent disturbance analysis of geological structure differences. Finally, it constructs a disturbance coefficient for disturbance analysis and uses this coefficient to improve the correction of sensor detection errors. In this way, it analyzes the differences in strain and environmental influence caused by geological inhomogeneity in the surrounding rock area, quantifies the impact of different geological structures on the deformation trend of the surrounding rock area, enhances the predictive ability of surrounding rock deformation using the quantitative analysis results, and uses the prediction results to correct sensor data, avoiding errors caused by inhomogeneous geology and thus improving the detection accuracy of surrounding rock deformation. Attached Figure Description

[0041] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This application provides a flowchart of the steps for a method to improve the accuracy of surrounding rock deformation detection;

[0043] Figure 2 This is a schematic diagram showing the location of the tunnel surrounding rock anchor bolts provided in this application;

[0044] Figure 3 A schematic diagram illustrating the acquisition of the perturbation coefficient provided in this application. Detailed Implementation

[0045] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method for improving the accuracy of surrounding rock deformation detection proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0046] Unless otherwise defined, terms such as “comprising,” “including,” or any other variations thereof are intended to cover a non-exclusive inclusion, such that a circuit structure, article, or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the article or device that includes said element. Furthermore, the term “and / or” as used herein includes any and all combinations of one or more of the associated listed items. All technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0047] The following description, in conjunction with the accompanying drawings, details a specific scheme for a method to improve the accuracy of surrounding rock deformation detection provided in this application.

[0048] This application provides an embodiment of a method for improving the accuracy of surrounding rock deformation detection. For details, please refer to [link to specific embodiments]. Figure 1 The method includes the following steps:

[0049] Step 1: Obtain strain data, ambient temperature, and ambient humidity at each monitoring point during each surrounding rock test.

[0050] Anchor bolts embedded with fiber Bragg grating (FBG) sensors are deployed at different locations on a cross-section of the tunnel's surrounding rock. Each anchor bolt is equipped with... Multiple monitoring points are used to achieve three-dimensional monitoring of the same location at different depths, and strain data is collected at each monitoring point. In this embodiment, the tunnel surrounding rock arch, left and right arch waists, and left and right arch shoulders are selected as the detection locations on the same cross-section. The value is set to 10. This application sorts the monitoring points at each location from smallest to largest depth. Then, the temperature and humidity of the surrounding rock environment are collected using the temperature and humidity sensors in the surrounding rock deformation monitoring system. All data are collected twice a day, and the collected data are normalized.

[0051] The schematic diagram of the location of the tunnel surrounding rock anchor bolts is shown below. Figure 2 As shown, Figure 2In the diagram, the dashed lines represent anchor bolts, the triangular sections represent monitoring points on the anchor bolts, the inner semicircle represents the tunnel, and the area between the inner and outer circles represents the surrounding rock region.

[0052] Step 2: Perform curve fitting on the strain data obtained from each monitoring point in each test and historical tests, analyze the shape characteristics of the fitted curves for all monitoring points in each test, and analyze the fluctuation of strain data of all monitoring points at the same location in each test to determine the geological difference coefficient for each test.

[0053] In surrounding rock deformation detection, anchor bolts equipped with fiber optic sensors are used to support the surrounding rock for real-time monitoring. The strain measured by the fiber optic sensors in the anchor bolts is used to detect the deformation. Under normal circumstances, when the surrounding rock is relatively uniformly distributed, the shear stress at each monitoring point on the anchor bolt initially decreases and then increases, with the shear stress decreasing closer to the anchor bolt's neutral point. Simultaneously, significant deformation displacement occurs in the initial stage of excavation, resulting in a large axial strain on the anchor bolt. The strain magnitude at each monitoring point also exhibits a certain regularity due to the shear stress variation. As construction progresses and the support system takes effect, the deformation displacement of the surrounding rock gradually decreases and stabilizes. When the geological structure within the surrounding rock area is unevenly distributed, the shear stress at various monitoring points on the anchor bolt exhibits poor disorder. Furthermore, when deformation and displacement occur in the surrounding rock, the deformation of different geological structures varies, resulting in a larger axial strain on the anchor bolt, and the strain under the influence of shear stress is also relatively disordered. Secondly, due to the mutual influence between different geological structures in the surrounding rock area, the final stable state will still fluctuate, representing a relatively stable state that requires a longer time to reach.

[0054] In the During the second test, the strain data from each monitoring point at the same location are used to construct the first strain data for that location according to the distribution order of the monitoring points. A strain sequence. A monitoring point is located before... The strain data are arranged in the order of detection to form a strain time series for the monitoring point. Using the elements in the strain time series of a monitoring point as the ordinate and the index of the element as the abscissa, a polynomial fitting is performed to output the fitted curve function. Polynomial fitting is a well-known technique and will not be described in detail here.

[0055] Based on this, a geological difference coefficient is calculated to measure the difference in geological homogeneity between different locations in the surrounding rock. Specifically:

[0056] First, for each detection, obtain the integral value formed by the fitted curve function and coordinate system of each monitoring point; obtain the integral value sequence composed of the integral values ​​of each monitoring point in each detection and all previous detections; obtain the first-order difference sequence of the integral value sequence, substitute the negative numbers of all elements in the first-order difference sequence into the sign function and average them, and then perform positive fusion with the integral value of each detection to obtain the strain trend of each monitoring point in each detection.

[0057] In this embodiment, the differences between variables are calculated using the difference value; the positive fusion of multiple variables is performed using the multiplication method, and the strain trend of the strain data of the nth monitoring point in the i-th detection is denoted as... Its formula is as follows: , , , They are the first The first monitoring point Second test, first Second test, first The integral value of the fitted curve for the second test. It is a symbolic function.

[0058] Furthermore, for each detection, the monitoring point corresponding to the minimum value of the strain sequence is obtained as the reference point; the percentage of negative differences in strain data between all adjacent monitoring points before the reference point is obtained, denoted as the first percentage; the percentage of positive differences in strain data between all adjacent monitoring points after the reference point is obtained, denoted as the second percentage; the average of the sum of the first percentage and the second percentage for all positions in each detection and all previous detections is used as the anchor strain trend measure for each detection, and the anchor strain trend measure for the i-th detection is denoted as... In this embodiment, the difference in strain data between adjacent monitoring points is specifically the difference between the strain data of the later monitoring point and the strain data of the earlier monitoring point.

[0059] Finally, the sum of the strain trend values ​​of all monitoring points in each test is obtained; the mean of the determination coefficients of the fitted curves of all monitoring points in each test is calculated and added to the anchor strain trend metric; the sum is negatively correlated and positively fused with the sum to obtain the geological difference coefficient for each test.

[0060] In this embodiment, an exponential method is used to perform negative correlation mapping on a single value, and a multiplication method is used to perform positive fusion of multiple variables. The geological difference coefficient of the i-th detection is denoted as... Its formula is as follows: In the formula, It is the first The strain trend of each monitoring point in the i-th detection. This is the total number of monitoring points. It is the first Measurement of anchor bolt strain trend in the second test. It is the first The mean of the coefficients of determination of the fitted curve function for all monitoring points during the test. It is an exponential function with the natural constant as its base.

[0061] It should be noted that since the coefficient of determination and the strain trend value range from 0 to 1, the average value of all monitoring points may be small. Therefore, an exponential function is used to enhance the difference.

[0062] It should be understood that in surrounding rock areas with uniform geological structure, the shear stress distribution on the anchor bolts shows a trend of first decreasing and then increasing. This directly leads to a similar trend in the axial strain of the anchor bolts, i.e., first decreasing and then increasing. Simultaneously, as construction progresses, the initial strain deformation is relatively large, gradually stabilizing later, but the overall strain is relatively small. Furthermore, because the surrounding rock area is relatively uniform, the stability of the surrounding rock area increases relatively with strain, and fluctuations decrease, resulting in a smaller geological difference coefficient, indicating a high degree of uniformity in the geological structure of the surrounding rock area. Conversely, the worse the uniformity of the geological structure of the surrounding rock area, the larger the corresponding geological difference coefficient.

[0063] Step 3: Map the ambient temperature of each test and historical tests onto a coordinate system, and connect the lines to obtain the temperature curve. The area corresponding to the temperature curve above the horizontal axis is taken as the freeze-thaw region. Based on the similarity between the strain data of each monitoring point in the non-freeze-thaw region and the ambient humidity in the corresponding freeze-thaw region, the similarity between the ambient humidity of each test and historical tests and the strain data of each monitoring point, and combined with the shape characteristics of the freeze-thaw region and the distribution differences of the strain data of the monitoring points in each test, determine the environmental correlation coefficient of each monitoring point in each test.

[0064] The uneven geological structure of the surrounding rock area not only affects the shear stress at each monitoring point on the anchor bolt, thus causing variations in strain data, but also leads to different degrees of environmental influence on different geological conditions. Normally, the selection of surrounding rock areas avoids areas with weak geological structures such as weak interlayers, soft rock, and fault fractures. For areas with more stable geological structures, the impact of the environment is less, ensuring the stability and safety of the surrounding rock. In other words, as the environment changes, the strain at different monitoring points is relatively similar and less affected by environmental changes. However, for areas with uneven geological structures, the varying degrees of environmental influence on different geological formations further increase the strain differences between monitoring points. Furthermore, monitoring points more affected by environmental fluctuations will experience further deterioration of their internal geological conditions due to natural changes such as humidity and freeze-thaw cycles. This deterioration leads to increased deformation of the surrounding rock, further amplifying the differences between monitoring points.

[0065] The ambient temperature and humidity data from each monitoring point, along with all previous measurements, are used to construct temperature and humidity sequences for each monitoring point, arranged in the order of the measurements. Addressing the significant freeze-thaw problem, which occurs when moisture within the surrounding rock freezes at low temperatures and melts at higher temperatures, damaging the internal structure and affecting the stability of the surrounding rock, a series of points are constructed using the elements in the temperature sequence as the ordinate and the element indices as the abscissa. Connecting adjacent points yields a temperature curve. A straight line is then plotted on the temperature curve. The area below is designated as the freeze-thaw zone, and the area above the line y=0 is designated as the non-freeze-thaw zone. The average strain of all strain data at each monitoring point within the corresponding segment of each non-freeze-thaw zone is calculated. The average strain of a monitoring point across all non-freeze-thaw zones is then used to construct a freeze-thaw strain sequence for that monitoring point, arranged in the order of the segments.

[0066] Based on the above analysis, an environmental correlation coefficient is calculated to measure the impact of environmental changes on the deformation of the surrounding rock at different locations. Specifically: First, for each monitoring point, the difference in strain data between each monitoring point and all other monitoring points in each test is calculated, and the average level of the difference in strain data between each monitoring point and all other monitoring points in each test is calculated. The ratio of the average level of each monitoring point in each test to the average level of the next test is used as the difference coefficient of each monitoring point in each test. In this embodiment, the difference in strain data is calculated by the difference between strain data, and the average level of multiple variables is calculated by the average value.

[0067] Furthermore, the correlation measure between the humidity sequence and the strain time series obtained for each monitoring point in each detection is obtained and recorded as the first correlation. The product between the area of ​​each freeze-thaw region and the average value of the ambient humidity in its corresponding horizontal axis interval is calculated, and the product is used as the freeze-thaw coefficient of the corresponding freeze-thaw region. The freeze-thaw coefficients obtained from each detection and all previous detections are combined to form a freeze-thaw sequence.

[0068] The correlation measure between the freeze-thaw strain sequence and the freeze-thaw sequence obtained for each monitoring point in each detection is obtained and denoted as the second correlation. The sum of the first correlation and the second correlation for each monitoring point in each detection is calculated. The negative correlation mapping result of the average difference coefficient obtained for each monitoring point in each detection and all previous detections is calculated and positively fused with the sum to obtain the environmental correlation coefficient of each monitoring point in each detection. In this embodiment, the correlation measure is calculated using the cosine similarity coefficient; the negative correlation mapping result of a variable is specifically the reciprocal of that variable; the positive fusion of multiple variables is calculated by multiplication.

[0069] It is understandable that the greater the environmental influence on the deformation of each monitoring point in the surrounding rock, the higher the correlation between its strain and humidity, and the greater the damage caused by freeze-thaw problems. The severity of freeze-thaw damage is related to temperature and humidity. The lower the temperature and the higher the humidity, the greater the freeze-thaw damage. At this time, the damage inside the surrounding rock after melting is more severe, and the corresponding strain is relatively greater. Secondly, the more severe the damage inside the surrounding rock, the more severe the deformation at the corresponding monitoring point will be, making the strain difference between this monitoring point and other monitoring points greater, so the corresponding environmental correlation coefficient is also greater.

[0070] Step 4: For each test, based on the geological difference coefficient and the environmental correlation coefficient of all monitoring points, determine the disturbance coefficient of each monitoring point, combine it with the regression model to obtain the strain prediction value for the next test, correct the strain data, and obtain the test results of surrounding rock deformation.

[0071] The heterogeneity of the surrounding rock geology can affect the accuracy of surrounding rock deformation detection, primarily due to the varying stability of different geological structures, which also exhibit different degrees of deterioration under environmental influences. Furthermore, the role of different geological structures in surrounding rock deformation and their sensitivity to environmental impacts are not only related to their own geological characteristics but also influenced by the overall geological distribution of the surrounding rock area. For example, for geological structures like soft rock or loess, which are highly susceptible to changes in humidity, if these structures are located at greater depths, their impact on shear stress at different monitoring points within the surrounding rock is relatively small, and they are also less affected by environmental fluctuations, thus causing relatively less interference with surrounding rock deformation detection. In such cases, we need to comprehensively consider the inherent characteristics of the geological structure and its sensitivity to environmental changes, determining the distribution of heterogeneous geological structures within the surrounding rock area based on this relationship, and assessing the impact of each location on surrounding rock deformation. This method can improve the accuracy of real-time monitoring and prediction of surrounding rock deformation, enhance detection precision, and effectively avoid safety hazards caused by the inability to accurately assess the development of surrounding rock deformation.

[0072] Based on the above analysis, the disturbance coefficient at each monitoring point is calculated to measure the degree of heterogeneity of the surrounding rock area and the impact of environmental changes on the deformation of the surrounding rock.

[0073] : , They are the first sequence Monitoring points during the second test The disturbance coefficient, , These are the geological difference coefficient of the i-th detection and the environmental correlation coefficient of the m-th monitoring point in the i-th detection, respectively. It was before The maximum value of the geological difference coefficient in this test. It is the first The maximum value among all geological difference coefficients at all monitoring points during the second test.

[0074] It should be noted that the initial perturbation coefficient is 1.

[0075] The diagram illustrating the acquisition of the disturbance coefficient is shown below. Figure 3 As shown.

[0076] It can be understood that the greater the heterogeneity of the geological structure of the surrounding rock area, and the greater the influence of the environment on this geological heterogeneity, the greater the impact of the corresponding location on the deformation of the surrounding rock area. In subsequent deformation prediction, the disturbance will be greater, thus resulting in a larger disturbance coefficient.

[0077] collection Each sample data point represents data from a single surrounding rock test, as described in this embodiment. The value is 360, which the implementer can choose according to the actual situation. Then, the detection data of each sample constitutes its feature vector. Select from 1 sample The remaining samples are used as validation samples, serving as training samples. The ratio of the perturbation coefficient to the maximum perturbation coefficient for each monitoring point in a sample is used as the weight of that monitoring point. The elements corresponding to ambient temperature and humidity are weighted to 1. The feature vectors of each sample are then weighted. The weighted feature vectors of the training samples form a feature matrix, where each row represents a feature vector of a single sample, and each column represents a single feature. Using this feature matrix as input, multiple linear regression is applied to output the predicted strain value for the next detection. Then, validation samples are used to adjust the hyperparameters, avoiding the risk of overfitting. Multiple linear regression is a well-known technique and will not be elaborated further.

[0078] Because the fiber Bragg grating sensor used in surrounding rock deformation detection is highly susceptible to environmental noise and temperature factors, leading to errors in the sensor results, a trained model can be used to correct these errors. Specifically, during subsequent surrounding rock deformation detection, real-time deformation is measured based on the data acquired at the detection moment. By comparing the current detected data with the predicted value, the error of the fiber Bragg grating sensor is corrected. Simultaneously, the acquired monitoring data is used to predict the surrounding rock deformation at the next moment.

[0079] In the formula: Is each monitoring point on the 1st The strain data corrected during the second test. These are the strain data of each monitoring point during the i-th measurement. It is the preset attenuation factor. It is the first Error during the second test , They are the first sequence The correction error for the second detection. In this embodiment, The value is set to 0.6. It should be noted that since there is no predicted value in the initial detection, the error of the initial detection is set to 0, and the correction error of the initial detection is set to 1.

[0080] Will Using the strain data of each sample as input, the Otsu thresholding algorithm is used to output a segmentation threshold. Otsu thresholding is a well-known technique and will not be elaborated further. In the... During the second test, the corrected strain data is acquired. If the corrected strain data is greater than the segmentation threshold, it is determined that there is surrounding rock deformation at the corresponding monitoring point; otherwise, there is no surrounding rock deformation.

[0081] It is understood that references to "one embodiment" or "some embodiments" in this specification mean that one or more embodiments of this application include the specific features, structures, or characteristics described in connection with that embodiment. Therefore, the appearance of phrases such as "in one embodiment," "in some embodiments," "in other embodiments," or "in still other embodiments" in different parts of this specification does not necessarily refer to the same embodiment, but rather means "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0082] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous. Moreover, the sequence numbers of the steps in the embodiments do not imply a specific order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments in this specification.

[0083] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for improving the accuracy of surrounding rock deformation detection, characterized in that, The method includes the following steps: Monitoring points are pre-defined at each location in the surrounding rock. Curve fitting is performed on the strain data obtained from each monitoring point in each test and historical tests. The shape differences of the fitted curves for each monitoring point in different tests are analyzed to determine the strain trend of each monitoring point in each test. The fluctuation of strain data from all monitoring points at the same location in each test is analyzed to obtain the anchor strain trend metric for each test. The cumulative sum of the strain trend of all monitoring points in each test is obtained. The mean of the determination coefficients of the fitted curves for all monitoring points in each test is calculated and added to the anchor strain trend metric. The sum is negatively correlated and positively fused with the cumulative sum to obtain the geological difference coefficient for each test. The ambient temperature of each test and historical tests is mapped onto a coordinate system, and the lines are connected sequentially to obtain a temperature curve. The area corresponding to the temperature curve above the horizontal axis is taken as the freeze-thaw region. Based on the similarity between the strain data of each monitoring point in the non-freeze-thaw region and the ambient humidity in the corresponding freeze-thaw region, the similarity between the ambient humidity of each test and historical tests and the strain data of each monitoring point, combined with the shape characteristics of the freeze-thaw region and the distribution differences of the strain data of the monitoring points in each test, the environmental correlation coefficient of each monitoring point in each test is determined. For each test, based on the geological difference coefficient and the environmental correlation coefficient of all monitoring points, the disturbance coefficient of each monitoring point is determined. Combined with the regression model, the strain prediction value for the next test is obtained. The strain data is then corrected to obtain the test results of the surrounding rock deformation.

2. The method for improving the accuracy of surrounding rock deformation detection as described in claim 1, characterized in that, Determining the strain trend for each monitoring point in each detection includes: First, for each detection, obtain the integral value formed by the fitted curve function and the coordinate system for each monitoring point; Obtain the integral value sequence of each monitoring point, which consists of the integral values ​​of each detection and all previous detections; obtain the first-order difference sequence of the integral value sequence, substitute the negative numbers of all elements in the first-order difference sequence into the sign function and average them, and then perform positive fusion with the integral value of each detection to obtain the strain trend quantity of each monitoring point in each detection.

3. The method for improving the accuracy of surrounding rock deformation detection as described in claim 1, characterized in that, The measure of anchor strain trend obtained for each test includes: The strain data of all monitoring points at the same location at each test is recorded as the strain sequence for each location; For each detection, the monitoring point corresponding to the minimum value of the strain sequence at each location is obtained as a reference point; The percentage of negative differences in strain data between all adjacent monitoring points preceding the reference point is recorded as the first percentage. The percentage of positive differences in strain data between all adjacent monitoring points following the reference point is recorded as the second percentage. The average of the sum of the first percentage and the second percentage for all locations in each test and all previous tests is used as the anchor strain trend measure for each test.

4. The method for improving the accuracy of surrounding rock deformation detection as described in claim 1, characterized in that, Determining the environmental correlation coefficient for each monitoring point in each detection includes: Based on the distribution of strain data at each monitoring point in the corresponding section of the non-freeze-thaw zone, the freeze-thaw strain sequence of each monitoring point is obtained; The distribution characteristics of ambient humidity in the corresponding sections of the freeze-thaw region were analyzed, and the freeze-thaw sequence was obtained by combining the area of ​​the freeze-thaw region. For each monitoring point, calculate the difference in strain data between each monitoring point and all other monitoring points in each test, and calculate the average level of the difference in strain data between each monitoring point and all other monitoring points in each test; use the ratio of the average level of each monitoring point in each test to the average level of the next test as the difference coefficient of each monitoring point in each test. The correlation measure between the humidity sequence and the strain time series obtained in each detection for each monitoring point is obtained and denoted as the first correlation. The correlation measure between the freeze-thaw strain sequence and the freeze-thaw sequence obtained in each detection for each monitoring point is obtained and denoted as the second correlation. The sum of the first correlation and the second correlation for each monitoring point in each detection is calculated. The negative correlation mapping result of the average difference coefficient obtained by each monitoring point in each detection and all previous detections is calculated and positively fused with the sum to obtain the environmental correlation coefficient of each monitoring point in each detection.

5. The method for improving the accuracy of surrounding rock deformation detection as described in claim 4, characterized in that, The obtained freeze-thaw strain sequence for each monitoring point includes: Calculate the average strain of all strain data at each monitoring point within the corresponding segment of each non-freeze-thaw region, and combine the average strain of each monitoring point across all non-freeze-thaw regions to form the freeze-thaw strain sequence for each monitoring point.

6. The method for improving the accuracy of surrounding rock deformation detection as described in claim 4, characterized in that, The steps for obtaining the freeze-thaw sequence are as follows: Calculate the product between the area of ​​each freeze-thaw region and the average value of the ambient humidity in its corresponding horizontal axis interval. Use this product as the freeze-thaw coefficient of the corresponding freeze-thaw region. Combine the freeze-thaw coefficients obtained from each detection and all previous detections to form a freeze-thaw sequence.

7. The method for improving the accuracy of surrounding rock deformation detection as described in claim 1, characterized in that, The determination of the disturbance coefficient for each monitoring point in each detection is specifically as follows: The formula for the disturbance coefficient is: In the formula, , They are the first sequence Monitoring points during the second test The disturbance coefficient, , These are the geological difference coefficient of the i-th detection and the environmental correlation coefficient of the m-th monitoring point in the i-th detection, respectively. It was before The maximum value of the geological difference coefficient in this test. It is the first The maximum value among all geological difference coefficients at all monitoring points during the second test.

8. The method for improving the accuracy of surrounding rock deformation detection as described in claim 1, characterized in that, The specific process of obtaining the strain prediction value for the next detection by combining the regression model is as follows: The ratio of the disturbance coefficient of each monitoring point in each detection to the maximum value of the disturbance coefficient of all monitoring points in each detection is used as the normalized result of the disturbance coefficient of each monitoring point in each detection; the normalized result is used as the weight of the strain data of each monitoring point in each detection, and the weights of the ambient temperature and ambient humidity in each detection are preset. A regression model is trained using a preset number of samples. Based on the data obtained from each monitoring point in each detection and historical detections, the strain prediction value for the next detection is obtained.

9. The method for improving the accuracy of surrounding rock deformation detection as described in claim 1, characterized in that, The process of correcting the strain data based on the error between the strain data detected in each test and the predicted strain value to obtain the detection results of the surrounding rock deformation includes: Each monitoring point will be on the [number]th The corrected strain data from the next test is denoted as Its formula is as follows: In the formula: Is each monitoring point on the 1st The strain data corrected during the second test. These are the strain data of each monitoring point during the i-th measurement. It is the preset attenuation factor. It is the first Error during the second test , They are the first sequence Correction error for the second test; A threshold segmentation algorithm is used to obtain a segmentation threshold from the strain data of a preset number of samples. When the corrected strain data is greater than the segmentation threshold, it is determined that there is no surrounding rock deformation at the corresponding monitoring point; otherwise, there is no surrounding rock deformation.

Images