A method for monitoring and early warning of surface subsidence at tunnel entrances

Abstract

The present invention provides a tunnel portal surface settlement monitoring and safety early warning method, which is applicable to the field of tunnel construction monitoring, including the arrangement of monitoring points, data collection and transmission, data processing and analysis, and safety early warning; the tunnel portal surface settlement monitoring and safety early warning method proposed in the present invention has the advantages of high safety and accuracy, and can be widely used in the field of tunnel construction monitoring.

Description

Technical Field

[0001] The invention relates to a tunnel opening surface settlement monitoring and safety early warning method, which is applicable to the field of tunnel construction monitoring. Background Art

[0002] As an important infrastructure construction, tunnel engineering is of great significance to promoting regional economic development and improving traffic conditions. However, during the tunnel construction process, due to the complex and changeable geological conditions, high construction technology requirements, and the unfavorable factors such as softness, fragmentation, and water content in the strata that the tunnel passes through, tunnel construction faces many safety risks, among which the surface settlement of the tunnel entrance is particularly prominent. The surface settlement of the tunnel entrance not only affects the stability and service life of the tunnel structure, but may also have a serious impact on the surrounding buildings, roads and environment, and even cause safety accidents. Therefore, effective monitoring and safety warning of the surface settlement of the tunnel entrance are crucial to ensuring construction safety and preventing accidents.

[0003] In the process of measuring the settlement of the tunnel entrance, the current common practice in the industry is to arrange multiple measuring lines above the tunnel entrance, perpendicular to the tunnel axis, and arrange monitoring points at a certain distance between each measuring line. The purpose of monitoring the settlement of the top of the tunnel entrance is achieved by measuring the settlement data of the monitoring points. At the same time, with the continuous development and popularization of computer technology, people in the industry have also begun to use some algorithms to establish models, analyze the monitoring data obtained, and obtain prediction values ​​based on the analysis rules to achieve the purpose of early warning.

[0004] However, practitioners currently ignore the special characteristics of tunnel portal settlement monitoring data. Tunnel portal settlement monitoring data often show high-frequency changes, especially when construction activities such as excavation and support are frequent in the tunnel portal area; tunnel portal settlement monitoring data will show obvious local settlement, this is because the tunnel portal area is a stress concentration area, which is very prone to abnormal local settlement; in view of the characteristics of high-frequency changes and local settlement in the above-mentioned tunnel portal settlement monitoring data, there is an urgent need for a tunnel portal surface settlement monitoring and safety early warning method, which can conduct targeted analysis based on the characteristics of high-frequency changes and local settlement. Summary of the invention

[0005] The purpose of the present invention is to solve the current urgent need for a tunnel vault settlement analysis method that can target the characteristics of high-frequency changes and local settlement, and propose a tunnel portal surface settlement monitoring and safety early warning method.

[0006] The purpose of the present invention can be achieved by adopting the following technical solutions:

[0007] A tunnel entrance surface settlement monitoring and safety early warning method, characterized in that the tunnel entrance surface settlement monitoring and safety early warning method comprises the following steps:

[0008] S101, arrangement of monitoring points;

[0009] The arrangement of the monitoring points includes arranging the surface settlement monitoring points at the tunnel entrance according to the topographic and geological conditions of the tunnel entrance. The surface settlement monitoring points are divided into multiple measuring lines and arranged perpendicular to the tunnel axis. The number of the surface settlement monitoring points is n, expressed as Di, i=1, 2, ..., n;

[0010] S102, data collection and transmission;

[0011] The data collection and transmission is to carry out measurement work at certain time intervals according to the surface subsidence monitoring points, obtain corresponding surface subsidence monitoring data, and transmit the monitoring data to the data processing center by wired or wireless means;

[0012] S103, data processing and analysis;

[0013] The processing and analysis of the data includes performing a data partition modeling analysis method on the surface subsidence monitoring data to obtain predicted data for surface subsidence monitoring;

[0014] The data processing and analysis also includes the steps of the data partition modeling and analysis method:

[0015] a) Obtain the minimum horizontal distance Hi and vertical distance Bi between each surface settlement monitoring point Di and the tunnel centerline, where i = 1, 2, ..., n;

[0016] b) The surface monitoring points are divided into two categories, Class I and Class II. The classification index is: when Hi corresponding to the surface settlement monitoring point i is greater than 2.5 times the design width of the tunnel section or the vertical distance Bi corresponding to it is greater than 2.5 times the design height of the tunnel section, the monitoring point is Class II, otherwise the monitoring point should be Class I. Finally, a total of k Class I surface settlement monitoring points are obtained, which are expressed as D 1 x , x = 1, 2, ..., k, the corresponding monitoring data set is represented by C 1 x , and at the same time, l type II surface settlement monitoring points are obtained, represented by D 2 y , y = 1, 2, ..., l, the corresponding monitoring data set is represented by C 2 y , where k+l=n;

[0017] c) Obtain a surface subsidence monitoring point i that needs to be analyzed,

[0018] d) When the classification result of the surface subsidence monitoring point i is Class I, a prediction model based on a long short-term memory neural network is used to construct a prediction model to obtain a prediction value y, and the long short-term memory neural network is used continuously to construct a prediction model to obtain prediction values ​​for a total of x times, and the x prediction values ​​are obtained. The mean and standard deviation of the x prediction values ​​are calculated, and a 90% confidence interval is calculated based on the calculation results of the mean and standard deviation and assuming that they obey a normal distribution, and finally a 90% confidence interval of the prediction value is obtained.

[0019] e) When the classification result of the surface subsidence monitoring point i is Class II, a prediction model is constructed using linear regression to obtain a prediction value,

[0020] f) Repeat the above steps c to e until all monitoring points are analyzed, and finally obtain the 95% confidence interval of the predicted value based on each Class I surface settlement monitoring point, and finally obtain the predicted value based on each Class II surface settlement monitoring point;

[0021] S104, safety warning;

[0022] The safety warning includes formulating warning standards 1 and 2 for Class I and Class II surface settlement monitoring points respectively based on the data processing and analysis and in accordance with the topographic and geological conditions of the tunnel entrance in combination with the classification standards;

[0023] The safety warning also includes that when the 95% confidence interval of the predicted value of a certain Class I surface subsidence monitoring point shows that it exceeds the warning standard 1, a safety warning is issued at the location of the Class I surface subsidence monitoring point; when the predicted value of a certain Class II surface subsidence monitoring point exceeds the warning standard 2, a safety warning is issued at the location of the Class II surface subsidence monitoring point;

[0024] Furthermore, in the above S103, the prediction model constructed by the long short-term memory neural network uses a non-fixed random seed during the weight initialization process;

[0025] Furthermore, in the above S103, the value of x is 100-1000;

[0026] Further, in the above S103, the method for obtaining the 90% confidence interval is: a) calculating the mean of the x predicted values, b) calculating the standard deviation of the x predicted values, c) determining the critical value of the standard normal distribution corresponding to the 90% confidence level, and d) calculating the confidence interval, the formula is formula (1),

[0027]

[0028] In the formula, is the mean, z is the critical value of the standard normal distribution, and s is the standard deviation.

[0029] The beneficial effects of the present invention are:

[0030] The beneficial effects of the present invention are as follows: according to the actual situation, it is proposed to divide all surface subsidence monitoring points into two categories, and at the same time adopt a data analysis modeling analysis method, especially introducing an uncertainty analysis method for Class I surface subsidence monitoring points, and using a 90% confidence interval to consider whether it exceeds the warning standard, which has the advantages of high safety and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 : A flow chart of a tunnel entrance surface settlement monitoring and safety early warning method of the present invention;

[0032] Figure 2 : Schematic diagram of the monitoring points and reference points of the tunnel surface settlement according to the embodiment of the present invention;

[0033] Figure 3 : A vertical line graph of the predicted value distribution of an embodiment of the present invention. DETAILED DESCRIPTION

[0034] The specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings; it should be understood that the specific embodiments given here are only used to illustrate and explain the present invention and cannot be used to limit the present invention.

[0035] The following is a specific embodiment of a tunnel entrance surface settlement monitoring and safety early warning method.

[0036] like Figure 1 FIG. 1 is a flow chart of a tunnel portal surface settlement monitoring and safety early warning method according to the present invention.

[0037] A tunnel entrance surface settlement monitoring and safety early warning method, characterized in that the tunnel entrance surface settlement monitoring and safety early warning method comprises the following steps:

[0038] S101, arrangement of monitoring points;

[0039] The arrangement of the monitoring points includes arranging the surface settlement monitoring points at the tunnel entrance according to the topographic and geological conditions of the tunnel entrance. The surface settlement monitoring points are divided into multiple measuring lines and arranged perpendicular to the tunnel axis. The number of the surface settlement monitoring points is n, which is expressed as D i , i=1,2,...,n;

[0040] In this embodiment, the supporting project is a tunnel construction project of the eighth division of the East Coast Rail Link project. The tunnel is 3595m long, with an entrance mileage of CH000+606 and an exit mileage of CH004+201. The line changes from an uphill slope of 3‰ from the entrance to a downhill slope of 6‰ at CH002+400, with a maximum burial depth of 335m and a minimum burial depth of 5m. The tunnel is designed as a ballasted composite lining structure with a standard line spacing of 4.2m. The Shuangwendan 1# tunnel is designed with an inclined shaft auxiliary construction and rescue channel with a length of 261m and an inclined shaft clearance section size of 5m*8m;

[0041] In this embodiment, in order to ensure the safety of tunnel opening excavation, 10 monitoring points and 1 reference point are arranged at the tunnel opening. Figure 1 , marked as D1~D 10 ;

[0042] S102, data collection and transmission;

[0043] The data collection and transmission is to carry out measurement work at certain time intervals according to the surface subsidence monitoring points, obtain corresponding surface subsidence monitoring data, and transmit the monitoring data to the data processing center by wired or wireless means;

[0044] In this embodiment, the surface settlement detection points are measured at regular time intervals. The time interval requirements are shown in Tables 1 and 2. In principle, a higher measurement frequency specified in Tables 1 and 2 is selected. At the same time, a total station instrument is used for measurement to obtain monitoring values, which are transmitted wirelessly to a data processing center.

[0045] Table 1 Monitoring measurement frequency determined by distance from the excavation working face

[0046] Distance between measuring section and excavation surface (m) Measurement frequency (0～1)B 2 times / day (1～2)B 1 time / day (2～5)B 1 time / 2~3 days >5B 1 time / 7 days

[0047] Note: B represents the tunnel excavation width

[0048] Table 2 Monitoring and measurement frequency determined by displacement speed

[0049] Displacement rate (mm / d) Measurement frequency ≥5 2 times / day 1～5 1 time / day 0.5～1 1 time / 2~3 days 0.2～0.5 1 time / 3 days <0.2 1 time / 7 days

[0050] S103, data processing and analysis;

[0051] The processing and analysis of the data includes performing a data partition modeling analysis method on the surface subsidence monitoring data to obtain predicted data for surface subsidence monitoring;

[0052] The data processing and analysis also includes the steps of the data partition modeling and analysis method:

[0053] a) Obtain the minimum horizontal distance Hi and vertical distance Bi between each surface settlement monitoring point Di and the tunnel centerline, where i = 1, 2, ..., n;

[0054] b) The surface monitoring points are divided into two categories, Class I and Class II. The classification index is: when Hi corresponding to the surface settlement monitoring point i is greater than 2.5 times the design width of the tunnel section or the vertical distance Bi corresponding to it is greater than 2.5 times the design height of the tunnel section, the monitoring point is Class II, otherwise the monitoring point should be Class I. Finally, a total of k Class I surface settlement monitoring points are obtained, which are expressed as D 1 x , x = 1, 2, ..., k, the corresponding monitoring data set is represented by C 1 x , and at the same time, l type II surface settlement monitoring points are obtained, represented by D 2 y , y = 1, 2, ..., l, the corresponding monitoring data set is represented by C 2 y , where k+l=n;

[0055] c) Obtain a surface subsidence monitoring point i that needs to be analyzed,

[0056] d) When the classification result of the surface subsidence monitoring point i is Class I, a prediction model based on a long short-term memory neural network is used to construct a prediction model to obtain a prediction value y, and the long short-term memory neural network is used continuously to construct a prediction model to obtain prediction values ​​for a total of x times, and the x prediction values ​​are obtained. The mean and standard deviation of the x prediction values ​​are calculated, and a 90% confidence interval is calculated based on the calculation results of the mean and standard deviation and assuming that they obey a normal distribution, and finally a 90% confidence interval of the prediction value is obtained.

[0057] e) When the classification result of the surface subsidence monitoring point i is Class II, a prediction model is constructed using linear regression to obtain a prediction value,

[0058] f) Repeat the above steps c to e until all monitoring points are analyzed, and finally obtain the 95% confidence interval of the predicted value based on each Class I surface settlement monitoring point, and finally obtain the predicted value based on each Class II surface settlement monitoring point;

[0059] Furthermore, in the above S103, the prediction model constructed by the long short-term memory neural network uses a non-fixed random seed during the weight initialization process;

[0060] Furthermore, in the above S103, the value of x is 100-1000;

[0061] Further, in the above S103, the method for obtaining the 90% confidence interval is: a) calculating the mean of the x predicted values, b) calculating the standard deviation of the x predicted values, c) determining the critical value of the standard normal distribution corresponding to the 90% confidence level, and d) calculating the confidence interval, the formula is formula (1),

[0062]

[0063] In the formula, is the mean, z is the critical value of the standard normal distribution, and s is the standard deviation;

[0064] In this embodiment, according to the classification method of the settlement monitoring points, the settlement monitoring points D1 to D 10 Divided into 2 categories, of which Class I points are D1, D2, D3, D9, D10, 5 points of each type, marked as D 1 x , x=1,2,...,5, where the type II points are D4, D5, D6, D7, D8, marked as; D 2 y, y=1,2,...,5; here the monitoring point D 1 1 Example study, the D 1 1 is a Class I monitoring point. Its existing monitoring data is obtained. A prediction model based on a long short-term memory neural network is used to obtain a prediction value. The long short-term memory neural network is used to construct a prediction model to obtain a prediction value 100 times in total. 100 prediction values ​​are obtained. The distribution of the 100 prediction values ​​is as follows: Figure 2 , the mean and standard deviation of the 100 predicted values ​​are calculated, and the 90% confidence interval is calculated based on the calculation results of the mean and standard deviation and assuming that they obey the normal distribution. The 90% confidence interval of the predicted value is finally obtained as [15.042, 15.449],

[0065] S104, safety warning;

[0066] The safety warning includes formulating warning standards 1 and 2 for Class I and Class II surface subsidence monitoring points respectively based on the data processing and analysis and according to the topographic and geological conditions of the tunnel entrance in combination with the classification standards; the safety warning also includes issuing a safety warning at the location of the Class I surface subsidence monitoring point when the 95% confidence interval of the predicted value of a Class I surface subsidence monitoring point shows that it exceeds the warning standard 1, and issuing a safety warning at the location of the Class II surface subsidence monitoring point when the predicted value of a Class II surface subsidence monitoring point exceeds the warning standard 2.

[0067] In this embodiment, the monitoring point D 11 is a Class I monitoring point, and the 90% confidence interval of the final predicted value is [15.042, 15.449]. On this basis, the predicted value is [15.042, 15.449], which is smaller than the warning standard of Class I monitoring points: the total settlement shall not be greater than 25 mm, then no warning is required.

[0068] In the above embodiments, the present invention discloses a method for monitoring and providing safety warning for surface subsidence at a tunnel entrance, including the arrangement of monitoring points, data collection and transmission, data processing and analysis, and safety warning; the method for monitoring and providing safety warning for surface subsidence at a tunnel entrance proposed in the present invention has the advantage of high safety and can be widely used in the field of tunnel construction monitoring.

[0069] The above description is a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for monitoring and early warning of tunnel entrance surface settlement, comprising the following steps: 1) arranging monitoring points, including arranging the tunnel entrance surface settlement monitoring points according to the topographic and geological conditions of the tunnel entrance, wherein the surface settlement monitoring points are divided into a plurality of measuring lines arranged perpendicular to the tunnel axis, and the number of the surface settlement monitoring points is n, expressed as D i , i = 1, 2, ..., n; 2) data acquisition and transmission, including carrying out measurement work at a certain time interval according to the surface subsidence monitoring points, obtaining corresponding surface subsidence monitoring data, and transmitting the monitoring data to the data processing center by wired or wireless means; 3) data processing and analysis; 4) safety warning; It is characterized in that, The processing and analysis of the data includes performing a data partition modeling analysis method on the surface subsidence monitoring data to obtain predicted data for surface subsidence monitoring; The data processing and analysis also includes the steps of the data partition modeling and analysis method: a) Obtain the location D of each surface subsidence monitoring point i Minimum lateral distance H from the tunnel centerline i and vertical distance B i ; b) The surface settlement monitoring points are divided into two categories, category I and category II. The classification standard for the two categories is: when the lateral distance H corresponding to the surface settlement monitoring point i i Greater than 2.5 times the tunnel section design width or its corresponding vertical distance B i When the height is greater than 2.5 times the design height of the tunnel section, the monitoring point is Class II, otherwise the monitoring point should be Class I. Finally, a total of k Class I surface settlement monitoring points are obtained, represented by D 1 x , x = 1, 2, ..., k, the corresponding monitoring data set is represented by C 1 x , and at the same time, l type II surface settlement monitoring points are obtained, represented by D 2 y , y = 1, 2, ..., l, the corresponding monitoring data set is represented by C 2 y , where k+l=n; c) Obtain a surface subsidence monitoring point i that needs to be analyzed, d) When the classification result of the surface subsidence monitoring point i is Class I, a prediction model based on a long short-term memory neural network is used to construct a prediction model to obtain a prediction value y. The long short-term memory neural network is used continuously to construct a prediction model to obtain a prediction value, and a total of x times are performed to obtain x prediction values. The mean and standard deviation of the x prediction values ​​are calculated. According to the calculation results of the mean and standard deviation and assuming that they obey a normal distribution, a 95% confidence interval is calculated, and finally a 95% confidence interval of the prediction value is obtained. e) When the classification result of the surface subsidence monitoring point i is Class II, a prediction model is constructed using linear regression to obtain a prediction value, f) Repeat the above steps c to e until all monitoring points are analyzed, and finally obtain the 95% confidence interval of the predicted value based on each Class I surface settlement monitoring point, and finally obtain the predicted value based on each Class II surface settlement monitoring point; The safety warning includes formulating warning standards 1 and 2 for Class I and Class II surface settlement monitoring points respectively based on the data processing and analysis and in accordance with the topographic and geological conditions of the tunnel entrance in combination with the classification standards; The safety warning also includes that when the 95% confidence interval of the predicted value of a Class I surface subsidence monitoring point exceeds the warning standard 1, a safety warning is issued at the location of the Class I surface subsidence monitoring point; when the predicted value of a Class II surface subsidence monitoring point exceeds the warning standard 2, a safety warning is issued at the location of the Class II surface subsidence monitoring point.

2. A tunnel portal surface settlement monitoring and safety early warning method according to claim 1, characterized in that: In the above 3), the prediction model constructed by the long short-term memory neural network uses an unfixed random seed during the weight initialization process.

3. A tunnel portal surface settlement monitoring and safety early warning method according to claim 1, characterized in that: In the above 3), the value of x is 100-1000.

4. A tunnel portal surface settlement monitoring and safety early warning method according to claim 1, characterized in that: In the above 3), the method for obtaining the 95% confidence interval is: a) calculating the mean of the x predicted values, b) calculating the standard deviation of the x predicted values, c) determining the critical value of the standard normal distribution corresponding to the 95% confidence level, and d) calculating the confidence interval, the formula is formula (1), In the formula, is the mean, z is the critical value of the standard normal distribution, and s is the standard deviation.

Images