A shield-formed tunnel morphology data processing system and method

Abstract

The present invention discloses a shield-formed tunnel morphology data processing system and method, which relates to the technical field of tunnel data processing. The shield-formed tunnel morphology data processing system and method specifically include the following steps: S1 Collect a number of cross-section data in the shield-formed tunnel and preprocess them; S2 Calculate the offset of the center of each segment; S3 Continuously measure and analyze the deformation trend; S4 Use a linear regression model to predict the change in offset; S5 Use the trained model to predict the final offset of the segment; Recording and analyzing the removed outliers helps to screen the abnormal factors of the segments in the tunnel, so as to timely discover and solve potential problems and prevent the problems from further deteriorating, thereby ensuring the safety and stability of the tunnel; Calculating the deformation and using time series analysis to analyze the deformation trend helps to timely discover the deformation trend and formulate subsequent detection plans. Through continuous monitoring, deformation problems can be timely discovered and processed to ensure the stability and safety of the tunnel.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel data processing, and specifically to a shield-formed tunnel morphology data processing system and method. Background Art

[0002] A shield tunneling method refers to an underground excavation method in which a shield machine is used to keep the excavation face stable, the formation soil is excavated under the protection of the shield shell, precast segments are assembled at its tail to form a tunnel structure, and finally a jack is used to support the assembled lining, and the shield is pushed forward by using its reaction force. The shield machine can automatically excavate, assemble segments, and push the shield, greatly saving manpower. And due to the automated operation of the shield machine, the construction speed is relatively fast, and the construction period can be shortened. The shield tunneling method can excavate a tunnel at one time, reducing the trouble of multiple excavations and support, and can be "tailor-made" according to the cross-sectional size, burial depth conditions, and basic conditions of the surrounding rock of the tunnel construction object, and is applicable to various complex geological conditions.

[0003] The segments in a shield-formed tunnel are the main structure of the tunnel lining, and their morphological data is crucial for the overall quality and safety of the tunnel. Publication No. CN105335576A discloses a shield-formed tunnel cross-section data processing method and system, which processes the data of disordered cross-section points to form a cross-section measurement result table, thereby simplifying the evaluation of tunnel engineering.

[0004] However, as shown in the above technology, it only solves the problem of intelligently processing a large amount of tunnel cross-section data to save manpower. Usually, data preprocessing is required for big data processing, such as cleaning, removing outliers, noise, etc., and then the missing data is supplemented by interpolation steps such as those in the above-mentioned comparative document. However, in fact, for the outliers in tunnel cross-section data, they are not necessarily caused by detection failures, because they are obtained by equipment scanning, and it is possible to scan abnormal conditions on the surface of the segments caused by unnoticed cracks, loosening of the segments, etc. In this case, the outliers actually have a certain warning effect. Therefore, it may not be appropriate to use common data processing logics; moreover, there are many reasons for the offset of the segments in the tunnel. The offset caused by construction factors may be one-time, but for some offsets caused by geological factors, they will be continuous. Therefore, it is necessary to continuously detect and predict them. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the present invention provides a shield-formed tunnel morphology data processing system and method, which solves the problems existing in the above prior art.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: A shield-formed tunnel morphology data processing method specifically includes the following steps:

[0007] S1 collects data of several cross-sections in a shield-constructed tunnel, classifies and sorts the data, cleans it by removing noise and outliers, interpolates missing data, and records the removed outliers for analysis of the reasons for anomalies to screen for abnormal factors of the segments in the tunnel;

[0008] S2 calculates the offset of the center of each segment and determines the actual distance of the offset, compares the offset with construction and geological information, and preliminarily analyzes the reasons for the offset;

[0009] S3 measures the segment diameter and calculates the roundness error; classifies and sorts the deformations, uses time series analysis to analyze the deformation trend, and formulates a subsequent detection plan, setting an upper limit on the number of measurements for continuous detection;

[0010] S4 combines time-domain factor analysis to analyze the offset law, uses a linear regression model to predict the change in the offset, selects a prediction model, and trains and validates the model using historical data;

[0011] S5 uses the trained model to predict the final offset of the segment, evaluates the reliability and uncertainty of the prediction results, and gives a confidence interval.

[0012] Preferably, the steps of step S1 specifically include:

[0013] S1.1 Data collection: Use high-precision measurement equipment to collect cross-section data at various key positions in the tunnel, and establish a cross-section data set U (u 1 ,u 2 ,...,u n ), where u represents the data of each cross-section, and u includes the segment center position O, diameter D, and thickness T. The segment center position O is represented by coordinates (X, Y, Z); the data collection frequency is determined according to the construction progress and geological conditions;

[0014] S1.2 Data classification and sorting: Classify the data according to types, including position data and dimension data; and sort each type of data in chronological order or by cross-section position;

[0015] S1.3 Data preprocessing: Clean the original data, remove noise and outliers, calculate the mean μ and standard deviation σ of the data set, and use linear interpolation or Lagrange interpolation to interpolate the missing data;

[0016] The formula for calculating the mean μ is: ;

[0017] where n is the number of data points in the data set, and u p is the p-th data point in U of the data set;

[0018] The formula for calculating the standard deviation σ is: 。

[0019] Preferably, when the acquisition device scans into the segment crack or there is a problem of local detachment of the segment, it will cause obvious data anomalies. Therefore, the identified outliers are recorded and marked while being excluded, which is used to judge the specific situation at the position of the outliers later. The specific steps are as follows:

[0020] a1 Outlier judgment: For each data point u in the data set U, calculate the Z value: Z = σu - μ; if |Z| > Z thresh , then the corresponding data point is regarded as an outlier;

[0021] a2 Outlier recording: The identified outliers are excluded from the original data set and recorded in the outlier record set A rec where the information recorded in A rec includes: the outlier itself, the acquisition time, the acquisition location, and the outlier type;

[0022] a3 Consecutive outlier judgment: Traverse A rec , and check whether there are data of N adj or more consecutive acquisition points that are all outliers. N adj is the set threshold of the number of adjacent acquisition points set to judge the situation of consecutive outliers;

[0023] a4 Warning trigger: If there are consecutive outliers, add the positions corresponding to the outliers to the warning set A warn , and trigger the warning mechanism to notify relevant personnel for handling;

[0024] a5 Subsequent processing: For the positions in A warn , if the segment displacement occurs at the same time, the outlier data is used as the analysis data for the subsequent cause of the segment displacement; then arrange for manual inspection or use other equipment to further confirm the position corresponding to the outlier.

[0025] Preferably, the condition for judging consecutive outliers is:

[0026] ;

[0027] where [p, q] represents the index range of consecutive acquisition points, and N represents the set of natural numbers, that is, the set of all positive integers; the condition p < q means that the starting index p is less than the ending index q, that is, [p, q] is a continuous data point interval, rather than a single point or a reverse interval; k is any integer between [p, q], and k ∈ A rec means that the data point corresponding to the index k exists in the outlier record set A rec and is a data point identified as an outlier; N adjN is the threshold value of the number of adjacent acquisition points, and q - p + 1 ≥ N adj indicates that the number of consecutive outliers from p to j is at least N adj .

[0028] Preferably, step S2 specifically includes:[[]]

[0029] S2.1 Calculate the offset: Use vector subtraction to calculate the offset ΔP = (ΔX, ΔY, ΔZ) of the center of each segment relative to the designed position; calculate the modulus length of the offset , which represents the actual distance of the offset;

[0030] S2.2 Preliminary analysis of the offset cause: Compare the offset with the construction records and geological exploration reports to preliminarily judge the offset cause; record the information of the time t, position (X′, Y′, Z′), and offset |ΔP| of the offset event.

[0031] Preferably, step S3 specifically includes:[[]]

[0032] S3.1 Diameter measurement and calculation: Measure the diameter of the segment at multiple angles to obtain the diameter array D j ={D j 1 , D j 2 ,..., D j m}, D j represents the diameter array of multiple acquisition points in the jth cross-section data u j in the dataset U; and calculate the roundness error , where max(D) is the maximum diameter value in D j , min(D) is the minimum diameter value in D j , and avg(D) is the average value of all diameter values in D j ;

[0033] S3.2 Deformation classification and sorting: Classify the deformation according to the degree and direction of deformation, including uniform deformation and local deformation; and sort the deformation according to the degree of deformation and the direction of deformation;

[0034] S3.3 Introduce the time series factor: Record the time t of each measurement to form a time series {t 1 , t 2 ,..., t n}, use the time series analysis method to analyze the change trend of deformation over time; Subsequent detection plan: Develop a subsequent detection plan, including detection time, detection location, and detection frequency; Set an upper limit M on the number of measurements for each offset position.

[0035] Preferably, step S4 specifically includes:[[]]

[0036] S4.1 Offset continuous detection: Set a threshold ε for judging whether there is a new offset in a certain measurement value of the offset position compared with the previous measurement. stop , if after a new detection is performed on a certain offset position, there is |ΔP new | - |ΔP old | < ε stop , it means that there is no new offset in the measurement position compared with the previous measurement. Then, stop the measurement at the position without a new offset amount, and continue the measurement at other positions until the measurement times reach the upper limit M; where, ΔP new represents the offset amount ΔP of the new measurement, and ΔP old represents the offset amount ΔP of the previous measurement.

[0037] S4.2 Offset law analysis: Analyze the offset law in combination with the time-domain factor, and use the linear regression model P′ = a + bt to predict the change trend of the offset amount over time, where P′ is the predicted offset amount, t is the time, a is the intercept, and b is the slope.

[0038] Preferably, step S5 specifically includes:

[0039] S5.1 Prediction model training: Use historical offset amount data to train and verify the model;

[0040] S5.2 Predict the final offset amount: Use the trained prediction model to predict the final offset amount P of the segment final ;

[0041] S5.3 Evaluate the reliability and uncertainty of the prediction result, and use the confidence interval [P lower , P upper to represent the range of the prediction result. The calculation steps of the confidence interval [P lower , P upper are as follows:

[0042] S5.3.1 Calculate the standard error of the predicted value:

[0043] The standard error of fit SEE of the model, that is, the standard deviation of the residuals, is expressed as: ;

[0044] where, H is the sample size, that is, the number of H samples collected at H time points, h = (1, 2,..., H); P h is the actually observed offset amount; is the predicted value of the model at time t h , that is, the predicted value at the h-th time point;

[0045] For the new time point t newThe standard error of prediction value SEP is expressed using the following approximate formula: ;

[0046] where, is the mean of the sample time;

[0047] Confidence interval: The standard error of the predicted value is used to calculate the confidence interval, expressed as: Confidence interval = (a + bt new ) ± 1.96 × SEP;

[0048] where z 0.025 ≈ 1.96 is the quantile of the standard normal distribution;

[0049] The calculated confidence interval is expressed as [P lower , P upper .

[0050] The present invention also discloses a shield - formed tunnel shape data processing system, including:

[0051] A data acquisition and pre - processing module, used for acquiring and pre - processing the cross - section data in the shield tunnel;

[0052] An offset calculation and analysis module, used for calculating the offset of the center of each segment relative to the designed position, determining the actual distance of the offset, comparing the offset with the construction records and geological exploration reports, and preliminarily analyzing the reasons for the offset;

[0053] A measurement and trend analysis module, used for measuring, calculating, classifying, and sorting the diameter of the segments, and introducing a time - series factor for trend analysis of deformation;

[0054] Analysis and prediction of offset rules, used for predicting the offset rules of the segments by combining the trend analysis of deformation and establishing a prediction model;

[0055] A prediction model training and evaluation module, used for training and evaluating the prediction model, and using the trained prediction model for formal offset prediction.

[0056] Preferably, it further includes:

[0057] A cloud platform, used for accessing and using system resources through the Internet;

[0058] A cloud storage module, used for storing all information in the acquisition and processing process;

[0059] A data visualization module, used for displaying the tunnel shape data in the form of graphs, images, or animations, including three - dimensional visualization of cross - section data, dynamic display of offsets, and chart analysis of deformation trends;

[0060] The data analysis and report generation module is used to generate a detailed report based on the analysis results, including data summary, analysis results, and recommended measures;

[0061] The system integration and interface module is used to integrate with the construction management system and geological exploration system, and provide open interfaces to allow third-party applications or services to access the system.

[0062] The present invention provides a shield-formed tunnel morphology data processing system and method. Compared with the prior art, it has the following beneficial effects:

[0063] 1. The shield-formed tunnel morphology data processing method records and analyzes the removed outliers, which helps to screen out the abnormal factors of the segments in the tunnel, so as to timely discover and solve potential problems, prevent the problems from further deteriorating, and thus ensure the safety and stability of the tunnel. Calculating the offset of the center of each segment and comparing it with the construction and geological information can preliminarily analyze the cause of the offset, which helps the construction team to timely adjust the construction plan and avoid or reduce the impact of the offset. Calculating the deformation and using time series analysis to analyze the deformation trend helps to timely discover the deformation trend, formulate subsequent detection plans, and through continuous monitoring, timely discover and handle the deformation problems to ensure the stability and safety of the tunnel.

[0064] 2. The shield-formed tunnel morphology data processing method, the outlier processing and continuous outlier judgment mechanism in the data preprocessing step effectively improve the quality and accuracy of the data. By recording and analyzing the outliers instead of directly removing them without analysis, potential problems such as segment cracks or local shedding can be timely discovered and handled, thus avoiding the impact of these problems on the safety and stability of the tunnel. This method also realizes the timely response and handling of the abnormal conditions of the segments in the tunnel through the triggering of the early warning mechanism and subsequent processing steps. This not only helps to ensure the safe operation of the tunnel, but also provides important reference information for the construction team so that they can adjust the construction plan and monitoring plan according to the actual situation.

[0065] 3. The shield-formed tunnel morphology data processing method can more accurately grasp the morphological changes of the tunnel and timely discover potential problems by accurately calculating the offset of the center of each segment and introducing the time series factor to analyze the deformation trend. At the same time, by deeply analyzing the offset law in combination with the time domain factor and using the linear regression model for prediction, the accuracy and reliability of the prediction are improved. In addition, this method also sets an upper limit on the number of measurements to avoid unnecessary resource waste, and introduces a confidence interval in the prediction result to provide more comprehensive information support for the construction team. In summary, this method shows obvious advantages in data processing, analysis and prediction, which helps to improve the safety and efficiency of tunnel construction and operation and reduce potential risks.

[0066] 4. The shield-formed tunnel shape data processing system integrates multiple functional modules such as data acquisition, preprocessing, analysis, prediction, and visualization, realizing the comprehensive management and efficient utilization of shield tunnel shape data. Through the cloud platform and cloud storage, remote access and storage of data are achieved, improving the sharing and security of data. The data visualization module makes the analysis results more intuitive and understandable, facilitating decision-making. In addition, the system integration and interface module enhances the scalability and compatibility of the system, facilitating seamless docking with other construction management and geological exploration systems, and improving the overall intelligent level of construction management. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 It is a schematic diagram of the overall process of the present invention;

[0068] Figure 2 It is a schematic diagram of the process of Embodiment 2 of the present invention;

[0069] Figure 3 It is a schematic diagram of the outlier judgment logic of Embodiment 3 of the present invention;

[0070] Figure 4 It is a schematic diagram of the process of Embodiment 3 of the present invention;

[0071] Figure 5 It is a block diagram of the system module of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0072] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0073] The present invention discloses a method for processing shield-formed tunnel shape data and provides the following three technical solutions:

[0074] Figure 1 The first implementation manner is shown: specifically including the following steps:

[0075] S1 Collect several cross-section data in the shield-formed tunnel, classify and sort the data and perform cleaning, remove noise and outliers, interpolate the missing data, and record the removed outliers for outlier cause analysis to screen the abnormal factors of the segments in the tunnel for outlier cause analysis to screen the abnormal factors of the segments in the tunnel;

[0076] S2 Calculate the offset of the center of each segment and determine the actual distance of the offset, compare the offset with the construction and geological information, and preliminarily analyze the cause of the offset;

[0077] S3 measures the segment diameter, calculates the roundness error; classifies and sorts the deformations, uses time series analysis to analyze the deformation trend, and formulates subsequent detection plans, setting an upper limit on the number of measurements for continuous detection;

[0078] S4 combines time domain factor analysis to analyze the offset law, uses a linear regression model to predict the change in the offset amount, and selects a prediction model, using historical data to train and validate the model;

[0079] S5 uses the trained model to predict the final offset of the segment, evaluates the reliability and uncertainty of the prediction results, and gives a confidence interval.

[0080] Recording and analyzing the removed outliers helps screen for abnormal factors of the segments in the tunnel, so as to timely discover and solve potential problems, prevent the problems from further deteriorating, and thus ensure the safety and stability of the tunnel. Calculating the offset of each segment center and comparing it with the construction and geological information can initially analyze the reasons for the offset, which helps the construction team timely adjust the construction plan and avoid or reduce the impact of the offset. Calculating the deformation and using time series analysis to analyze the deformation trend helps timely discover the deformation trend, formulate subsequent detection plans, and through continuous monitoring, timely discover and handle deformation problems to ensure the stability and safety of the tunnel.

[0081] Figures 2 - 3 The second implementation manner is shown, and the main difference from the first implementation manner is that: the steps of step S1 specifically include:

[0082] S1.1 Data collection: Use high-precision measurement equipment to collect cross-section data at various key positions in the tunnel, and establish a cross-section data set U (u 1 , u 2 ,..., u n ), where u represents the data of each cross-section, and u includes the segment center position O, diameter D, and thickness T. The segment center position O is represented by coordinates (X, Y, Z); the data collection frequency is determined according to the construction progress and geological conditions to ensure the timeliness and representativeness of the data;

[0083] S1.2 Data classification and sorting: Classify the data according to types, including position data and dimension data; and sort each type of data in chronological order or cross-section position for subsequent analysis;

[0084] S1.3 Data preprocessing: Clean the original data, remove noise and outliers, such as removing outliers using the 3σ principle, calculate the mean μ and standard deviation σ of the data set, and use linear interpolation or Lagrange interpolation method to interpolate the missing data;

[0085] The calculation formula for the mean μ is: ;

[0086] Among them, n is the number of data points in the dataset, and u p is the p-th data point in U of the dataset;

[0087] The calculation formula for the standard deviation σ is: .

[0088] When the acquisition device scans into the segment crack or there is a problem of local detachment of the segment, it will cause obvious data anomalies. Therefore, the identified outliers are removed and recorded and marked at the same time for subsequent judgment of the specific situation at the outlier, and the specific steps are as follows:

[0089] a1 Outlier judgment: For each data point u in the dataset U, calculate the Z value: Z = σu - μ; if |Z| > Z thresh ; (Z thresh is the set threshold, such as 3), then the corresponding data point is regarded as an outlier;

[0090] a2 Outlier recording: Remove the identified outliers from the original dataset and record them in the outlier record set A rec , and the information recorded in A rec includes: the outlier itself, the acquisition time, the acquisition location, and the outlier type (such as data anomalies caused by cracks);

[0091] a3 Consecutive outlier judgment: Traverse A rec , and check whether there are data of N adj or more consecutive acquisition points that are all outliers. N adj is the set threshold for the number of adjacent acquisition points to judge the situation of consecutive outliers;

[0092] a4 Warning trigger: If there are consecutive outliers, add the positions corresponding to the outliers to the warning set A warn , and trigger the warning mechanism to notify relevant personnel for processing;

[0093] a5 Follow-up processing: For the positions in A warn , if the segment displacement occurs at the same time, the outlier data is used as the analysis data for the subsequent cause of the segment displacement; then arrange for manual inspection or use other equipment to further confirm the position corresponding to the outlier, such as checking whether there are inconspicuous cracks.

[0094] The condition for judging consecutive outliers is:

[0095] ;

[0096] Among them, [p, q] represents the index range of consecutive acquisition points, and N represents the set of natural numbers, that is, the set of all positive integers; the condition p < q means that the starting index p is less than the ending index q, that is, [p, q] is a continuous data point interval, rather than a single point or a reverse interval; k is any integer between [p, q], and k ∈ A rec indicates that the data point corresponding to the index k exists in the outlier record set A rec and is a data point identified as an outlier; N adj is the threshold for the number of adjacent acquisition points, and q - p + 1 ≥ N adj indicates that the number of consecutive outliers from p to j is at least N adj .

[0097] The outlier processing and continuous outlier judgment mechanism in the data preprocessing step effectively improve the quality and accuracy of the data. By recording and analyzing outliers instead of directly removing them without analysis, potential problems such as segment cracks or local detachment can be discovered and processed in a timely manner, thus avoiding the impact of these problems on the safety and stability of the tunnel. This method also realizes the timely response and processing of segment anomalies in the tunnel through the triggering of the early warning mechanism and subsequent processing steps. This not only helps to ensure the safe operation of the tunnel, but also provides important reference information for the construction team so that they can adjust the construction plan and monitoring plan according to the actual situation.

[0098] Figure 4 shows the third implementation manner, and the main difference from the first implementation manner is that: step S2 specifically includes:

[0099] S2.1 Calculate the offset: Use vector subtraction to calculate the offset ΔP = (ΔX, ΔY, ΔZ) of each segment center relative to the designed position; calculate the modulus length of the offset , indicating the actual distance of the offset;

[0100] S2.2 Preliminary analysis of the offset cause: Compare the offset with the construction records and geological exploration reports to preliminarily judge the offset cause; record the information of the time t, position (X′, Y′, Z′), and offset |ΔP| of the offset event.

[0101] Step S3 specifically includes:

[0102] S3.1 Diameter measurement and calculation: Measure the diameter of the segment at multiple angles to obtain the diameter array D j ={D j 1 , D j 2 ,..., D j m}}j Denote the diameter array of multiple acquisition points in the j-th cross-section data u in the dataset U; and calculate the roundness error j ; where max(D) is the maximum diameter value in D , and min(D) is the minimum diameter value in D j , and avg(D) is the average value of all diameter values in D j ; j The average value of all diameter values in

[0103] S3.2 Deformation classification and sorting: Classify the deformation according to the degree and direction of deformation, including uniform deformation and local deformation; and sort the deformation, including by degree of deformation and direction of deformation;

[0104] S3.3 Introduce the time series factor: Record the time t of each measurement to form a time series {t 1 , t 2 ,..., t n}, and use time series analysis methods, such as the ARIMA model, to analyze the change trend of deformation over time; Subsequent detection plan: Develop a subsequent detection plan, including detection time, detection location, and detection frequency; Set an upper limit M on the number of measurements for each offset position.

[0105] Step S4 specifically includes:

[0106] S4.1 Continuous detection of offset: Set a threshold ε for determining whether there is a new offset in a certain measurement value of the offset position compared to the previous measurement stop , if after a new detection is performed at a certain offset position, there is |ΔP new | - |ΔP old | < ε stop , it means that there is no new offset at the measurement position compared to the previous measurement, so stop measuring at the position with no new offset, and continue measuring at other positions until the number of measurements reaches the upper limit M; where, ΔP new represents the offset ΔP of the new measurement, and ΔP old represents the offset ΔP of the previous measurement;

[0107] S4.2 Analysis of offset pattern: Analyze the offset pattern in combination with the time domain factor, and use the linear regression model P′ = a + bt to predict the change trend of the offset over time, where P′ is the predicted offset, t is the time, a is the intercept, and b is the slope.

[0108] Step S5 specifically includes:

[0109] S5.1 Prediction model training: Use historical offset data to train and validate the model to ensure the accuracy and reliability of the model;

[0110] S5.2 Predict the final offset: Use the trained prediction model to predict the final offset P of the segment final ;

[0111] S5.3 Evaluate the reliability and uncertainty of the prediction results. Use the confidence interval [P lower , P upper to represent the range of the prediction results. The calculation steps of the confidence interval [P lower , P upper are as follows:

[0112] S5.3.1 Calculate the standard error of the predicted value:

[0113] The standard error of the estimate SEE of the model fit, that is, the standard deviation of the residuals, is expressed as: ;

[0114] where H is the sample size, that is, the number of H samples collected at H time points, h = (1, 2,..., H); P h is the actually observed offset; is the predicted value of the model at time t h , that is, the predicted value at the h-th time point;

[0115] For the standard error of the predicted value SEP at the new time point t new , it is expressed using the following approximate formula: ;

[0116] where is the mean of the sample times;

[0117] Confidence interval: Use the standard error of the predicted value to calculate the confidence interval, which is expressed as: Confidence interval = (a + bt new ) ± 1.96 × SEP;

[0118] where z 0.025 ≈ 1.96 is the quantile of the standard normal distribution, indicating a confidence level of 95%. Since 1 - 0.95 = 0.05 and 0.05 / 2 = 0.025, z 0.025 is selected;

[0119] The calculated confidence interval is expressed as [P lower , P upper .

[0120] By precisely calculating the offset of each segment center and introducing a time series factor to analyze the deformation trend, this method can more accurately grasp the morphological changes of the tunnel and timely detect potential problems. At the same time, by deeply analyzing the offset law in combination with the time domain factor and using a linear regression model for prediction, the accuracy and reliability of the prediction are improved. In addition, this method also sets an upper limit on the number of measurements to avoid unnecessary resource waste, and introduces a confidence interval in the prediction results to provide more comprehensive information support for the construction team. In summary, this method shows obvious advantages in data processing, analysis and prediction, which helps to improve the safety and efficiency of tunnel construction and operation and reduce potential risks.

[0121] Refer to Figure 5 , the present invention also discloses a shield-formed tunnel morphology data processing system, including:

[0122] A data acquisition and preprocessing module for collecting and preprocessing the cross-section data inside the shield tunnel;

[0123] An offset calculation and analysis module for calculating the offset of each segment center relative to the design position, determining the actual distance of the offset, comparing the offset with the construction records and geological exploration reports, and preliminarily analyzing the reasons for the offset;

[0124] A measurement and trend analysis module for measuring, calculating, classifying and sorting the diameter of the segments, and introducing a time series factor for trend analysis of the deformation;

[0125] Analysis and prediction of offset law for predicting the offset law of segments in combination with the trend analysis of deformation and establishing a prediction model;

[0126] A prediction model training and evaluation module for training and evaluating the prediction model and using the trained prediction model for formal offset prediction;

[0127] A cloud platform for accessing and using system resources through the Internet;

[0128] A cloud storage module for storing all information in the acquisition and processing process;

[0129] A data visualization module for displaying the tunnel morphology data in the form of graphics, images or animations, including 3D visualization of cross-section data, dynamic display of offsets, and chart analysis of deformation trends;

[0130] A data analysis and report generation module for generating a detailed report according to the analysis results, including data summary, analysis results, and recommended measures;

[0131] A system integration and interface module for integrating with the construction management system and geological exploration system, and providing an open interface to allow third-party applications or services to access the system.

[0132] The system integrates multiple functional modules such as data acquisition, preprocessing, analysis, prediction, and visualization, achieving comprehensive management and efficient utilization of shield tunnel morphology data. Through the cloud platform and cloud storage, remote access and storage of data are realized, improving data sharing and security. The data visualization module makes the analysis results more intuitive and understandable, facilitating decision-making. In addition, the system integration and interface module enhance the scalability and compatibility of the system, facilitating seamless docking with other construction management and geological exploration systems, and improving the overall intelligent level of construction management.

[0133] Meanwhile, the content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0134] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device.

[0135] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A shield-forming tunnel morphology data processing method, characterized in that: The specific steps include: S1 collects data from several sections in the shield-shaped tunnel, sorts and cleans the data, removes noise and outliers, interpolates missing data, and records the removed outliers for abnormal cause analysis to screen abnormal factors of the segments in the tunnel; when the acquisition equipment scans the cracks in the segments, or when the segments are partially detached, the data will be obviously abnormal, so the identified outliers are removed and recorded and marked for subsequent judgment of the specific situation of the outliers. The specific steps are as follows: a1 Outlier judgment: Data set U is a set of cross-sectional data collected at key locations in the tunnel using high-precision measuring equipment. For each data point u in data set U, calculate the Z value: Z=σu−μ; if |Z|>Z thresh , the corresponding data point is regarded as an outlier, where μ and σ are the mean and standard deviation of each data point u in the data set U, respectively. thresh is the abnormal threshold; a2 Outlier record: Remove the identified outliers from the original data set and record them in the outlier record set A rec A rec The information recorded in the includes: the outlier itself, the time of collection, the location of collection, and the type of outlier; a3 Continuous outlier judgment: traverse A rec , check whether there are consecutive N adj The data of the collection points with or above N are all abnormal values. adj The threshold value of the number of adjacent collection points is set to determine the situation of continuous outliers; a4 Warning trigger: If there are continuous outliers, the position corresponding to the outliers is added to the warning set A warn , and trigger the early warning mechanism to notify relevant personnel to handle it; a5 Subsequent processing: For A warn If the segment displacement occurs at the same time, the abnormal value data will be used as the analysis data of the cause of the subsequent segment displacement; then arrange manual inspection or use other equipment to further confirm the position corresponding to the abnormal value; S2 calculates the offset of each segment center and determines the actual distance of the offset, compares the offset with construction and geological information, and preliminarily analyzes the cause of the offset; S3 measures the diameter of the pipe segment and calculates the roundness error; classifies and sorts the deformation, introduces the timing factor, and records the time t of each measurement to form a time series {t1, t2, ..., t n }, use time series analysis methods to analyze the change trend of deformation over time, formulate subsequent inspection plans, and set an upper limit on the number of measurements for continuous inspection; S4 combines the time domain factor analysis with the offset law, uses the linear regression model to predict the offset change, selects the prediction model, and uses historical data to train and verify the model; S5 uses the trained model to predict the final displacement of the segment, evaluates the reliability and uncertainty of the prediction results, and provides a confidence interval.

2. A shield-forming tunnel morphology data processing method according to claim 1, characterized in that: The steps of step S1 specifically include: S1.1 Data collection: Use high-precision measurement equipment to collect cross-sectional data at key locations in the tunnel and establish a cross-sectional data set U (u1, u2, ..., u n ), where u represents the data of each section, and u includes the center position O, diameter D, and thickness T of the segment. The center position O of the segment is represented by the coordinates (X, Y, Z); the frequency of data collection is determined according to the construction progress and geological conditions; S1.2 Data classification and sorting: Classify data by type, including location data and size data; and sort each type of data by time sequence or cross-sectional position; S1.3 Data preprocessing: clean the raw data, remove noise and outliers, calculate the mean μ and standard deviation σ of the data set, and use linear interpolation or Lagrange interpolation to interpolate missing data; The calculation formula for the average value μ is: ; Where n is the number of data points in the dataset, u p is the pth data point in U in the dataset; The calculation formula of standard deviation σ is: .

3. A shield-forming tunnel morphology data processing method according to claim 1, characterized in that: The conditions for judging continuous outliers are: ; Where [p, q] represents the index range of continuous collection points, N represents the set of natural numbers, that is, the set of all positive integers; the condition p < q means that the starting index p is less than the ending index q, that is, [p, q] is a continuous data point interval, not a single point or reverse interval; k is any integer between [p, q], and k∈A rec Indicates that the data point corresponding to index k exists in the outlier record set A rec In the example, there is a data point identified as an outlier; N adj is the threshold of the number of adjacent collection points, q−p+1≥N adj Indicates that the number of consecutive outliers from p to j is at least N adj .

4. A shield-forming tunnel morphology data processing method according to claim 1, characterized in that: Step S2 specifically includes: S2.1 Calculate the offset: Use vector subtraction to calculate the offset of each segment center relative to the design position ΔP = (ΔX, ΔY, ΔZ); calculate the modulus length of the offset , represents the actual distance of the offset; S2.2 Preliminary analysis of the cause of the offset: Compare the offset with the construction records and geological survey reports to preliminarily determine the cause of the offset; record the time t, position (X′, Y′, Z′) and offset |ΔP| of the offset event.

5. A shield-forming tunnel morphology data processing method according to claim 4, characterized in that: Step S3 specifically includes: S3.1 Diameter measurement and calculation: Measure the diameter of the segment at multiple angles to obtain the diameter array D j = {D j 1 , D j 2 , ..., D j m }, D j Represents the j-th section data u in the data set U j The diameter array of multiple collected points in the circle; and calculate the roundness error , where max(D) is D j The maximum diameter value in, min (D) is D j The minimum diameter value in, avg(D) is D j The average value of all diameter values ​​in; S3.2 Classification and sorting of deformation: Classify deformation according to the degree and direction of deformation, including uniform deformation and local deformation; and sort deformation, including by degree and direction of deformation; S3.3 Subsequent inspection plan: Develop a subsequent inspection plan, including inspection time, inspection location, and inspection frequency; set an upper limit M on the number of measurements for each offset position.

6. A shield-forming tunnel morphology data processing method according to claim 5, characterized in that: Step S4 specifically includes: S4.1 Continuous offset detection: Set the threshold ε to determine whether a certain measurement value of the offset position has a new offset compared to the previous measurement stop , if a new detection is performed at a certain offset position, there is |ΔP new |-|ΔP old | < ε stop , it means that there is no new offset in the measurement position compared with the last measurement, so the measurement is stopped at the position without new offset, and the measurement is continued at other positions until the number of measurements reaches the upper limit M; where ΔP new Indicates the offset of the new measurement ΔP, ΔP old Indicates the offset ΔP of the previous measurement; S4.2 Analysis of offset rules: The offset rules are analyzed in combination with time domain factors, and the linear regression model P′= a +bt is used to predict the trend of offset changes over time, where P′ is the predicted offset, t is time, a is the intercept, and b is the slope.

7. A shield-forming tunnel morphology data processing method according to claim 6, characterized in that: Step S5 specifically includes: S5.1 Prediction model training: Use historical offset data to train and validate the model; S5.2 Predict the final offset: Use the trained prediction model to predict the final offset P of the segment. final ; S5.3 Assess the reliability and uncertainty of predictions using confidence intervals [P lower , P upper ] indicates the range of the prediction results, and the confidence interval [P lower , P upper The calculation steps of ] are: S5.3.1 Calculate the standard error of the prediction: The standard error SEE of the model fitting, that is, the standard deviation of the residuals, is expressed as: ; Where H is the sample size, that is, the number of H samples collected at H time points, h = (1, 2, ..., H); P h is the actual observed offset; is the model at time t h The predicted value at time , that is, the predicted value at the hth time point; For the new time point t new The standard error of the predicted value SEP is expressed using the following approximate formula: ; in, is the mean of the sample time; Confidence interval: The confidence interval is calculated using the standard error of the predicted value, expressed as: Confidence interval = (a + bt new ) ±1.96×SEP; Among them, z 0.025 ≈ 1.96 is the quantile of the standard normal distribution; The calculated confidence interval is expressed as [P lower , P upper ].

8. A shield-shaped tunnel morphology data processing system, used to implement the shield-shaped tunnel morphology data processing method according to any one of claims 1 to 7, characterized in that: include: Data acquisition and preprocessing module, used to acquire and preprocess the cross-section data in the shield tunnel; The offset calculation and analysis module is used to calculate the offset of each segment center relative to the designed position, determine the actual distance of the offset, compare the offset with the construction records and geological survey reports, and preliminarily analyze the cause of the offset; The measurement and trend analysis module is used to measure, calculate, classify and sort the diameter of the segments, and introduce time series factors to perform deformation trend analysis; Analysis and prediction of displacement patterns, used to predict segment displacement patterns in combination with deformation trend analysis and establish a prediction model; The prediction model training and evaluation module is used to train and evaluate the prediction model and use the trained prediction model to perform formal offset prediction.

9. A shield-forming tunnel morphology data processing system according to claim 8, characterized in that: Also includes: Cloud platforms, used to access and use system resources via the Internet; Cloud storage module, used to store all information of the collection and processing process; Data visualization module, used to display tunnel morphology data in the form of graphics, images or animations, including 3D visualization of cross-section data, dynamic display of offsets and graphical analysis of deformation trends; Data analysis and report generation module, used to generate detailed reports based on analysis results, including data summary, analysis results, and recommended measures; The system integration and interface module is used to integrate with the construction management system and geological survey system, and provides an open interface to allow third-party applications or services to access the system.

Images