Method for evaluating structural safety by using tunnel health monitoring data

Abstract

The invention discloses a method for evaluating structural safety by using tunnel health monitoring data, which comprises the following steps of: acquiring monitoring data to obtain real-time monitoring data; performing abnormal data judgment; correcting the abnormal data; calculating a threshold value of a single index; performing single index evaluation by adopting a membership function; tunnel health comprehensive evaluation grading is carried out; calculating a correction coefficient by adopting an entropy weight method; and tunnel health comprehensive evaluation is carried out step by step. According to the invention, rapid and quantitative evaluation of tunnel structure safety is realized, the problems of low manual inspection efficiency, poor quality and incapability of mastering the stress state of the tunnel in the traditional tunnel operation and maintenance process are solved, the purpose of rapidly identifying the most unfavorable stress area of the tunnel can be met, and the tunnel operation and maintenance efficiency is improved. The purpose of providing a large amount of data to support formulation of disease treatment measures when diseases occur can be achieved, the idea of intelligent operation and maintenance is met, the automation degree is high, practicability is high, and the obvious application and popularization value is achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of safety assessment of tunnels and underground projects, and particularly relates to a method for evaluating the structural safety by using tunnel health monitoring data. Background Art

[0002] In the daily operation and maintenance process of tunnel projects, at the current stage, manual inspections and visual inspections are mainly carried out, and only a small number of apparent diseases and some internal diseases that have developed seriously enough to cause changes in the appearance of the tunnel can be found. This operation and maintenance method greatly restricts the discovery of diseases in the tunnel. Most diseases can only be discovered after they occur. Even if repairs are carried out after the diseases occur, the diseases have already occurred, and the structural damage is irreversible, resulting in a reduction in the structural bearing capacity and safety reserve. In severe cases, structural safety problems may even occur due to untimely treatment of diseases, further affecting traffic safety. In addition, after diseases occur in the tunnel structure, most of the repair and treatment measures are based on experience, lacking an understanding of the stress conditions of the tunnel structure itself, resulting in insufficient pertinence of the formulated disease treatment measures. The treatment measures for some parts with less serious diseases are too strong, causing waste, and the treatment measures for some parts with more serious diseases are insufficient and do not eliminate potential hazards. Therefore, it is very necessary to monitor indicators such as external loads, stress and strain, overall deformation, local deformation, and durability of the structure through sensors buried in the health monitoring system during the tunnel operation and maintenance process, master the safety status of the tunnel structure, and pre-emptively eliminate diseases.

[0003] In the process of using health monitoring data to evaluate structural safety, the selection of monitoring indicators and the determination of evaluation levels are the key. It is necessary to comprehensively master the monitoring data of the tunnel external environment, structural response, overall deformation, local deformation, structural durability, etc. through reasonable selection of monitoring indicators in order to have a comprehensive understanding of the safety status of the tunnel structure. The determination of the evaluation level is to quickly and accurately locate the weak positions of the tunnel through different-level evaluation results during the actual operation and maintenance process, and quickly guide the formulation of the operation and maintenance plan. In addition, the tunnel structural safety evaluation is divided into single-index evaluation and comprehensive analysis evaluation. Single-index evaluation is the basis of all evaluations, that is, using tunnel structure calculations, code provisions, material ultimate strength, etc. as division indicators to evaluate single monitoring indicators, which can be used as the structural safety evaluation result of a single measuring point position and also as the basis for subsequent comprehensive analysis evaluation. Comprehensive analysis and evaluation uses the results of single-index evaluation to evaluate the overall structural safety of the tunnel.

[0004] Currently, most of the existing tunnel structural safety evaluation methods mainly focus on displacement evaluation, which can qualitatively analyze the tunnel structural safety but cannot accurately and quantitatively analyze. Therefore, it is necessary to develop a new analysis and evaluation method to accurately analyze the tunnel structural safety. Summary of the Invention

[0005] The present invention is proposed to solve the problems existing in the prior art, and its purpose is to provide a method for evaluating the structural safety using tunnel health monitoring data.

[0006] The technical solution of the present invention is: a method for evaluating the structural safety using tunnel health monitoring data, including the following steps:

[0007] A. Collect monitoring data to obtain real-time monitoring data;

[0008] B. Judge abnormal data;

[0009] C. Perform correction processing on abnormal data;

[0010] D. Calculate the boundary values of single indicators;

[0011] E. Evaluate single indicators using membership functions;

[0012] F. Conduct comprehensive evaluation and grading of tunnel health;

[0013] G. Calculate the correction coefficient using the entropy weight method;

[0014] H. Conduct comprehensive evaluation of tunnel health step by step.

[0015] Furthermore, in step A, the monitoring data is collected to obtain real-time monitoring data, and the specific process is as follows:

[0016] First, adopt automated monitoring means to continuously collect the monitoring indicators of tunnel health monitoring sensors without interruption;

[0017] Then, obtain the above real-time monitoring data.

[0018] Furthermore, in step B, the abnormal data is judged, and the specific process is as follows:

[0019] First, judge the collected real-time monitoring data through the Grubbs method, Dixon method, and empirical judgment method;

[0020] Then, when the judgment result of any of the above methods is abnormal, this frame of monitoring data is regarded as abnormal data.

[0021] Furthermore, in step C, the correction processing is performed on the abnormal data, and the specific process is as follows:

[0022] First, use the Lagrange interpolation method to correct single-frame abnormal data;

[0023] Then, for continuous abnormal data, use the method of elimination for processing.

[0024] Furthermore, in step D, the boundary values of single indicators are calculated, and the specific process is as follows:

[0025] First, the calculation of the limit value of a single index is the basis for the evaluation of a single index;

[0026] Then, for each monitoring index, three limit values are set from the four aspects of structural calculation and analysis, code provisions, material limits, and on-site implementation experience;

[0027] Finally, four health level intervals are divided according to the three limit values.

[0028] Furthermore, in step E, the membership function is used for the evaluation of a single index, and the specific process is as follows:

[0029] First, the membership function is used to represent the membership degree of the factor set to the evaluation set;

[0030] Then, the health degree of each single monitoring factor in the actual project is basically linearly related to its monitoring value, and a membership function combining triangular distribution and trapezoidal distribution is adopted;

[0031] Finally, combined with the limit value of the single detection index, the single monitoring index is evaluated to obtain a score.

[0032] Furthermore, in step F, the comprehensive evaluation and grading of tunnel health are carried out, and the specific process is as follows:

[0033] First, the comprehensive evaluation of tunnel health is divided into six levels;

[0034] Then, the above six levels include the health evaluation of a single measurement point, the health evaluation of a single monitoring index of a single cross-section, the health evaluation of a certain type of monitoring index of a single cross-section, the health evaluation of the monitoring cross-section, the health evaluation of the same structural type, and the overall health evaluation of the tunnel.

[0035] Furthermore, in step G, the entropy weight method is used to calculate the correction coefficient, and the specific process is as follows:

[0036] First, the historical monitoring data of multiple measurement points are placed into the same matrix for matrix normalization;

[0037] Then, the weight is calculated and recorded as the correction coefficient for subsequent calculations.

[0038] Furthermore, in step H, the comprehensive evaluation of tunnel health is carried out step by step, and the specific process is as follows:

[0039] First, from the perspective of engineering practice application, the comprehensive evaluation of tunnel health is established as follows. Highlight the most critical factors. From the safety perspective, the benchmark should be the single evaluation value with the most unfavorable safety condition among the disease indexes;

[0040] Then, considering the superposition effect of multiple factors and the impact of other indexes on the overall health, the correction term should be subtracted from the benchmark value;

[0041] Subsequently, combined with the correction coefficient, the evaluation score of this level is obtained;

[0042] Finally, the evaluation score of the previous level is continuously calculated using the evaluation score of this level to complete the comprehensive evaluation of the tunnel health.

[0043] The beneficial effects of the present invention are as follows:

[0044] Under the environment of tunnel automatic health monitoring, based on the health monitoring data and the structural characteristics of the tunnel itself, the present invention realizes the rapid and quantitative evaluation of the tunnel structure safety, solves the problems of low efficiency and poor quality of manual inspection during the traditional tunnel operation and maintenance process, and is unable to master the stress state of the tunnel, can meet the purpose of quickly identifying the most unfavorable stress area of the tunnel, can achieve the purpose of providing a large amount of data to support the formulation of disease treatment measures when diseases occur, conforms to the concept of intelligent operation and maintenance, has very important application value in the field of tunnel operation and maintenance, is an indispensable part of realizing the intelligent operation and maintenance of the whole life cycle of the tunnel, and promotes the optimization of the tunnel design theory at the same time.

[0045] The present invention has a high degree of automation and strong practicability, and has obvious popularization and application value. Description of the Drawings

[0046] Figure 1 is the method flow chart of the present invention;

[0047] Figure 2 is the critical value G(α,n) of the Grubbs criterion in the present invention;

[0048] Figure 3 is the calculation method of the Dixon method in the present invention;

[0049] Figure 4 is the schematic diagram of the membership function in the present invention. Detailed Embodiments

[0050] Hereinafter, the present invention will be described in detail with reference to the drawings and embodiments:

[0051] As Figures 1 to 4 shown, a method for evaluating the structural safety using tunnel health monitoring data includes the following steps:

[0052] A. Perform monitoring data collection to obtain real-time monitoring data;

[0053] B. Perform abnormal data judgment;

[0054] C. Perform correction processing on the abnormal data;

[0055] D. Perform calculation of the limit value of single index;

[0056] E. Use membership functions for single - index evaluation;

[0057] F. Conduct comprehensive evaluation and grading of tunnel health;

[0058] G. Calculate the correction coefficient using the entropy weight method;

[0059] H. Conduct comprehensive evaluation of tunnel health step by step.

[0060] In step A, monitor data is collected to obtain real - time monitoring data. The specific process is as follows:

[0061] First, use automated monitoring means to continuously collect the monitoring indicators of tunnel health monitoring sensors without interruption;

[0062] Then, obtain the above - mentioned real - time monitoring data.

[0063] In step B, abnormal data is judged. The specific process is as follows:

[0064] First, judge the collected real - time monitoring data through the Grubbs method, Dixon method, and empirical judgment method;

[0065] Then, when the judgment result of any of the above methods is abnormal, this frame of monitoring data is regarded as abnormal data.

[0066] In step C, correction processing is carried out for abnormal data. The specific process is as follows:

[0067] First, use the Lagrange interpolation method to correct single - frame abnormal data;

[0068] Then, for continuous abnormal data, use the method of elimination for processing.

[0069] In step D, the boundary values of single - index are calculated. The specific process is as follows:

[0070] First, the calculation of the boundary values of single - index is the basis for single - index evaluation;

[0071] Then, three boundary values are set for each monitoring index from four aspects: structural calculation and analysis, code provisions, material limits, and on - site implementation experience;

[0072] Finally, divide into four health - level intervals according to the three boundary values.

[0073] In step E, membership functions are used for single - index evaluation. The specific process is as follows:

[0074] First, membership functions are used to represent the membership degree of the factor set to the evaluation set;

[0075] Then, the health degree of each single monitoring factor in the actual project basically conforms to a linear relationship with its monitoring value, and a membership function combining triangular distribution and trapezoidal distribution is adopted;

[0076] Finally, combined with the limit values of single detection indexes, the single monitoring indexes are evaluated to obtain scores.

[0077] Step F conducts a comprehensive evaluation and grading of tunnel health, and the specific process is as follows:

[0078] First, the comprehensive evaluation of tunnel health is divided into six levels;

[0079] Then, the above six levels include the health evaluation of single measurement points, the health evaluation of a certain monitoring index of a single cross-section, the health evaluation of a certain type of monitoring index of a single cross-section, the health evaluation of monitoring cross-sections, the health evaluation of the same structural type, and the overall health evaluation of the tunnel.

[0080] Step G calculates the correction coefficient by using the entropy weight method, and the specific process is as follows:

[0081] First, the historical monitoring data of multiple measurement points are placed into the same matrix for matrix normalization;

[0082] Then, the weights are calculated and recorded as correction coefficients for subsequent calculations.

[0083] Step H conducts a comprehensive evaluation of tunnel health level by level, and the specific process is as follows:

[0084] First, from the perspective of engineering practice applications, the comprehensive evaluation of tunnel health is established as follows. Highlight the most critical factors. From a safety perspective, the single evaluation value with the most unfavorable safety condition among each disease index should be used as the benchmark;

[0085] Then, considering the superposition effect of multiple factors and the impact of other indexes on the overall health, the correction term should be subtracted from the benchmark value;

[0086] Next, combined with the correction coefficient, the evaluation score of this level is obtained;

[0087] Finally, the evaluation score of this level is used to continue calculating the evaluation score of the previous level to complete the comprehensive evaluation of tunnel health.

[0088] Specifically, the Grubbs method in step B is as follows:

[0089] First, the Grubbs criterion based on the mean and standard deviation is a typical outlier detection method based on parametric statistics. The Grubbs criterion is premised on the W normal distribution, with rigorous theory and convenient use.

[0090] Then, calculate the average value as follows:

[0091]

[0092] After that, calculate the standard deviation as follows:

[0093]

[0094] After that, calculate each Gi value as follows:

[0095]

[0096] Finally, compare the Gi value with the Grubbs critical value G(α,n). If Gi > G(α,n), it is determined as an outlier. The Grubbs method critical value G(α,n) is as Figure 2 shown.

[0097] Specifically, the Dixon method in step B is as follows:

[0098] First, the Dixon method believes that the abnormal data should be the maximum and minimum data. Therefore, its basic method is to sort the data by size and check whether the maximum and minimum data are abnormal data.

[0099] Then, sort the data, X1 ≤ X2 ≤ X3… ≤ X n .

[0100] After that, calculate the Dixon statistic.

[0101] After that, look up the table to get the Dixon test critical value D(α, n).

[0102] Finally, make a determination. If D > D’ and D > D(α,n), it is judged as abnormal data. Or if D’ > D and D < D(α,n), after removing the outlier, if the number of removed outliers is less than the critical value (manually specified) compared to n, the outlier removal ends. Or if the determination conditions are not met, the outlier removal ends.

[0103] If the end conditions are not met, recalculate using the new data series after removing the outliers. The calculation methods of D (for testing high - end outliers) and D’ (for testing low - end outliers) in the Dixon method are as Figure 3 .

[0104] Specifically, the empirical judgment method in step B is as follows:

[0105] First, judge the abnormal temperature data

[0106] For the sensor temperatures on the same segment, exclude the data of the sensor itself and calculate the average. When the measured value exceeds 1.5 times the range after excluding itself and is greater than 2℃, it is judged as abnormal data.

[0107] The correction method is to use the average value of the sensor temperatures on this segment after removing the abnormal data to replace the outlier.

[0108] Then, judge other abnormal data

[0109] Collect the monitoring data of the same type of sensors on the same segment ring in the recent half month. After excluding the historical abnormal values, calculate the current offset value / range of each sensor. When the offset value / range of a certain sensor is more than 4 times the average value of the offset values / ranges of other sensors, it is regarded as an abnormal value.

[0110] Specifically, in step C, the Lagrange interpolation method is used to correct the single-frame abnormal data, as follows:

[0111] First, construct the Lagrange basis function Li(t), that is

[0112]

[0113] Then, the interpolation polynomial is

[0114] Finally, calculate the monitoring correction value at * the moment of t

[0115] Specifically, in step D, based on the corrected data, calculate the boundary values of single indicators, as follows:[[]]

[0116] First, the calculation of the boundary values of single indicators is the basis for the evaluation of single indicators;

[0117] Then, three boundary values are set for each monitoring indicator from the four aspects of structural calculation analysis, code provisions, material limits, and on-site implementation experience;

[0118] Finally, divide into four health level intervals according to the three boundary values.

[0119] Among them, the value corresponding to the first-level boundary value at the initial stage of the project can be selected as the standard combination value of the load. As the operation time increases, the representative value of the measured value can be selected.

[0120] Among them, the value corresponding to the second-level boundary value at the initial stage of the project can be selected as the basic combination / structural importance coefficient of the load. As the operation time increases, the frequent value of the measured value can be selected.

[0121] Among them, the value corresponding to the third-level boundary value at the initial stage of the project can be selected as the basic combination value of the load. As the operation time increases, the limit value of the measured value can be selected.

[0122] Specifically, in step E, the membership function is used to evaluate single indicators, and the specific process is as follows:[[]]

[0123] First, the membership function is used to represent the membership degree of the factor set to the evaluation set;

[0124] Then, the health degree of each individual monitoring factor in the actual project is basically linearly related to its monitoring value, and a membership function combining triangular distribution and trapezoidal distribution is adopted;

[0125] Finally, the individual monitoring indicators are evaluated in combination with the boundary values of the individual detection indicators to obtain scores.

[0126] As Figure 4 shown. In the figure, k1 - k3 are the boundary values for dividing 4 health level intervals, and the corresponding state boundary values also correspond to the first - level boundary value, the second - level boundary value, and the third - level boundary value respectively. Combining Figure 4 , the membership functions for each health interval are respectively:

[0127]

[0128]

[0129] Specifically, step G calculates the correction coefficient using the entropy weight method as follows:

[0130] First, the multi - measurement - point monitoring historical data is placed into the same matrix, and matrix normalization is performed in the way of (X - Xmin) / (Xmax - Xmin)+0.001. The normalized matrix is N = [x ij n×m .

[0131] Then, calculate the proportion of the i - th sample value under the j - th indicator to this indicator, that is:

[0132]

[0133] Next, calculate the entropy value of the j - th indicator (column), that is:

[0134]

[0135] Among them, usually take

[0136] Next, calculate the difference coefficient of the j - th indicator (column), that is:

[0137] d j =1 - e j .

[0138] Next, calculate the weight of the j - th indicator (column), that is:

[0139]

[0140] Finally, record ω j as the correction coefficient for subsequent calculations.

[0141] ​Specifically, in step H, a comprehensive tunnel health assessment is carried out step by step as follows:

[0142] First, from the perspective of engineering practice applications, the comprehensive tunnel health assessment is established as follows: First, highlight the most critical factors. From a safety perspective, the benchmark F should be the single evaluation value with the most unfavorable safety condition among all disease indexes i (that is, select Min from the values of each disease index). Considering the superposition effect of multiple factors and the impact of other indexes on the overall health, a correction term o(F i ) should be subtracted from the benchmark value, that is, F = F i - o(F i ). Ensure the convergence of data, ensure that F = F i - o(F i ) ≥ 1 holds (because from the initial mathematical relationship, the single F value is greater than 1). When approaching 1, the value of the correction term should be smaller. Tunnel comprehensive evaluation method where Fi = min{Ma, b, n}; F j is different from F i ; F j0 is the value when there is no disease in the jth item.

[0143] Specifically, in step H, a comprehensive tunnel health assessment is carried out step by step as follows:

[0144] From the perspective of engineering practice applications, the comprehensive tunnel health assessment is established as follows:

[0145] First, highlight the most critical factors. From a safety perspective, the benchmark F should be the single evaluation value with the most unfavorable safety condition among all disease indexes i (that is, select Min from the values of each disease index).

[0146] Then, considering the superposition effect of multiple factors and the impact of other indexes on the overall health, a correction term o(F i ) should be subtracted from the benchmark value, that is, F = F i - o(F i ).

[0147] Next, ensure the convergence of data, ensure that F = F i - o(F i ) ≥ 1 holds (because from the initial mathematical relationship, the single F value is greater than 1). When approaching 1, the value of the correction term should be smaller.

[0148] Next, the tunnel comprehensive evaluation method is as follows:

[0149]

[0150] Among them, F i= min{M a,b,n}; F j is different from each item of F i ; F j0 is the value when the j-th item has no disease.

[0151] Finally, as the evaluation score of this level, after the calculation is completed, the evaluation score of this level is used to continue calculating the evaluation score of the upper level.

[0152] Under the environment of tunnel automated health monitoring, based on health monitoring data and the structural characteristics of the tunnel itself, the present invention realizes a rapid and quantitative evaluation of the tunnel structural safety, solves the problems of low efficiency and poor quality of manual inspection during the traditional tunnel operation and maintenance process, and the inability to master the stress state of the tunnel, can meet the purpose of quickly identifying the most unfavorable stress area of the tunnel, can achieve the purpose of providing a large amount of data to support the formulation of disease treatment measures when diseases occur, conforms to the concept of intelligent operation and maintenance, has very important application value in the field of tunnel operation and maintenance, is an indispensable part of realizing the intelligent operation and maintenance of the whole life cycle of the tunnel, and at the same time promotes the optimization of the tunnel design theory.

[0153] The present invention has a high degree of automation and strong practicability, and has obvious promotion and application value.

Claims

1. A method for evaluating the structural safety using tunnel health monitoring data, characterized in that: It includes the following steps: A. Conduct monitoring data collection to obtain real-time monitoring data; B. Judge abnormal data; C. Perform correction processing on abnormal data; D. Calculate the boundary values of single indicators; E. Evaluate single indicators using membership functions; F. Conduct comprehensive evaluation and grading of tunnel health; G. Calculate the correction coefficient using the entropy weight method; H. Conduct comprehensive evaluation of tunnel health step by step.

2. The method for evaluating structural safety using tunnel health monitoring data according to claim 1, wherein: In step A, monitoring data collection is carried out to obtain real-time monitoring data. The specific process is as follows: First, adopt automated monitoring means to continuously collect the monitoring indicators of tunnel health monitoring sensors without interruption; Then, obtain the above real-time monitoring data.

3. A method for evaluating structural safety using tunnel health monitoring data according to claim 1, characterized in that: In step B, abnormal data is judged. The specific process is as follows: First, judge the collected real-time monitoring data through the Grubbs method, Dixon method, and empirical judgment method; Then, when the judgment result of any of the above methods is abnormal, this frame of monitoring data is regarded as abnormal data.

4. A method for evaluating structural safety using tunnel health monitoring data according to claim 1, characterized in that: In step C, correction processing is performed on abnormal data. The specific process is as follows: First, use the Lagrange interpolation method to correct single-frame abnormal data; Then, for continuous abnormal data, use the method of elimination for processing.

5. A method for evaluating structural safety using tunnel health monitoring data according to claim 1, characterized in that: In step D, the boundary values of single indicators are calculated. The specific process is as follows: First, the calculation of the boundary values of single indicators is the basis for the evaluation of single indicators; Then, three boundary values are set for each monitoring indicator from four aspects: structural calculation and analysis, code provisions, material limits, and on-site implementation experience; Finally, four health level intervals are divided according to the three boundary values.

6. A method for evaluating structural safety using tunnel health monitoring data according to claim 1, characterized in that: In step E, membership functions are used to evaluate single indicators. The specific process is as follows: First, membership functions are used to represent the membership degree of the factor set to the evaluation set; Then, the health degree of each single monitoring factor in actual engineering is basically linearly related to its monitoring value, and a membership function combining triangular distribution and trapezoidal distribution is adopted; Finally, combined with the boundary values of single detection indicators, single monitoring indicators are evaluated to obtain scores.

7. A method for evaluating structural safety using tunnel health monitoring data according to claim 1, characterized in that: In step F, comprehensive evaluation and grading of tunnel health are carried out. The specific process is as follows: First, the comprehensive evaluation of tunnel health is divided into six levels; Then, the above six levels include single-measurement point health evaluation, health evaluation of a certain monitoring indicator of a single cross-section, health evaluation of a certain type of monitoring indicator of a single cross-section, health evaluation of the monitoring cross-section, health evaluation of the same structural type, and overall tunnel health evaluation.

8. A method for evaluating structural safety using tunnel health monitoring data according to claim 1, characterized in that: In step G, the entropy weight method is used to calculate the correction coefficient. The specific process is as follows: First, place the multi-measurement point monitoring historical data into the same matrix and perform matrix normalization; Then, calculate the weights and record the weights as correction coefficients for subsequent calculations.

9. A method for evaluating structural safety using tunnel health monitoring data according to claim 1, characterized in that: In step H, comprehensive evaluation of tunnel health is carried out step by step. The specific process is as follows: First, from the perspective of engineering practice applications, the comprehensive evaluation of tunnel health is established as follows. Highlight the most critical factors. From a safety perspective, the benchmark should be the single evaluation value with the most unfavorable safety condition among various disease indicators; Then, considering the superposition effect of multiple factors and the impact of other indicators on the overall health, a correction term should be subtracted from the benchmark value; Next, combined with the correction coefficient, obtain the evaluation score of this level; Finally, use the evaluation score of this level to continue calculating the evaluation score of the previous level to complete the comprehensive evaluation of tunnel health.

Images