A method for predicting lateral settlement of overburden soil in subway tunnel excavation construction

Abstract

The present application relates to a kind of subway tunnel excavation construction overburden lateral settlement prediction method, belong to subway tunnel technical field.The present application can effectively utilize the actual monitoring data in subway tunnel excavation construction, carry out the correction of Peck theory formula and the training of neural network, realize the lateral settlement prediction of overburden of mountain city subway tunnel excavation construction, provide basic data for the demarcation of subway excavation construction settlement influence range and support structure design.

Description

Technical Field

[0001] This invention belongs to the field of subway tunnel technology and relates to a method for predicting the lateral settlement of the overlying soil layer during subway tunnel excavation. Background Technology

[0002] Subway tunnel excavation inevitably leads to ground subsidence, which can adversely affect roads, bridges, or buildings located above the tunnel. Therefore, the extent of subsidence caused by the cut-and-cover construction is a key factor to consider during subway tunnel construction in mountainous cities, in order to develop effective support structures.

[0003] Current methods for settlement analysis and prediction in subway tunnel excavation mainly include empirical formula methods, numerical analysis methods, and machine learning methods. Empirical formula methods are primarily based on research into the impact of subway tunnel construction on the overlying soil layer. Empirical formulas are constructed and parameters are fitted and corrected using measured data to obtain settlement prediction formulas applicable to certain conditions. This method can define the range of overlying soil layer settlement influence, but its applicability is difficult to control when geological conditions change. Numerical analysis methods use finite element method software to simulate subway tunnel construction, considering the impact of different construction processes on overlying soil layer settlement. This method can also define the range of settlement influence for subway tunnel excavation, but it requires prior knowledge of the overlying soil layer parameters around the subway tunnel. This limits the application of the method in practical engineering, and the method requires the construction of a finite element model of the tunnel, which makes it difficult to meet the timeliness requirements of engineering. The machine learning method uses intelligent algorithms such as neural networks and deep models, combined with measured data, to predict the settlement of the overlying soil layer during subway tunnel construction. This method is highly portable and can make full use of the monitoring data accumulated during tunnel construction, and has great application prospects. However, since the settlement data that can be tested during construction is usually the data of various measuring points on the ground surface, the machine learning method can only obtain the settlement prediction of certain specific points, and it is difficult to delineate the settlement range of the overlying soil layer.

[0004] Therefore, in order to fully leverage the advantages of the empirical formula method in delineating the settlement range of the overlying soil layer and the machine learning method in making dynamic predictions using actual monitoring data, it is necessary to further explore the organic combination of the two methods to form a method for predicting the transverse surface settlement of the overlying soil layer in subway tunnel excavation based on empirical formulas and machine learning methods, so as to provide a basis for delineating the settlement impact range and designing the support structure during subway excavation. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a method for predicting the lateral settlement of the overlying soil layer during the underground excavation of subway tunnels, so as to give full play to the advantages of the empirical formula method in delineating the settlement range of the overlying soil layer and the machine learning method in making dynamic predictions using actual monitoring data, and to provide basic data for delineating the settlement impact range and designing the support structure during the underground excavation of subways.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for predicting the lateral settlement of the overlying soil layer during the cut-and-cover construction of subway tunnels, the method comprising the following steps:

[0008] S1: The expression for Peck's formula is defined as follows:

[0009]

[0010] S max It is a parameter related to the formation volume loss rate and the maximum ground settlement; i is the width coefficient of the settlement trough;

[0011] When obtaining cross-sectional monitoring data of a subway line through testing, and performing parameter fitting, S... max Set the maximum settlement value when x = 0 to ensure that the fitting formula can correctly reflect the maximum settlement value of the subway cross section; transform equation (1) into equation (2) to fit parameter i;

[0012]

[0013] If the cross-sectional distance x and the settlement S are known... x Then S max =4.41, calculate i based on the data at each point. 2 The mean of the values ​​is used as the final estimate, where points where x is 0 are not included in the calculation. 2 =260.71; the fitted Peck formula is

[0014] S2: Obtain the neural network training set; use the geometric conditions, geological conditions, and tunnel excavation speed of each tunnel cross-section as input variables; use the Peck formula parameters S of each tunnel cross-section obtained from monitoring. max and i are used as the model output; the training dataset consists of actual monitored settlement data from constructed tunnel sections;

[0015] S3: Training and validating the neural network model;

[0016] S4: Using a trained and validated neural network model, calculate the Peck formula parameters S for subway tunnel excavation. maxThe identification of i; for a tunnel section about to be excavated by cut-and-cover method, input the geometric conditions, geological conditions, and tunnel excavation speed of the cross section to obtain the Peck formula parameters S of the tunnel cross section to be constructed. max and i;

[0017] S5: Identify the Peck formula parameters S max Substituting i into formula (1), we obtain the Peck formula for the cross section of the tunnel to be constructed, so as to determine the settlement influence range and settlement distribution law of the cross section of the tunnel to be constructed.

[0018] Optionally, the geometric conditions include tunnel height, tunnel span, and tunnel depth, and the geological conditions include elastic modulus, internal friction angle, cohesion, and Poisson's ratio.

[0019] The beneficial effects of this invention are as follows: This invention can effectively utilize actual monitoring data in the construction of subway tunnels by cutting and excavating to correct Peck's theoretical formula and train neural networks, thereby realizing the prediction of lateral settlement of the overlying soil layer in the construction of subway tunnels in mountainous cities by cutting and excavating, and providing basic data for the delineation of the settlement influence range and the design of support structures in the construction of subway tunnels by cutting and excavating.

[0020] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0021] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:

[0022] Figure 1 This is a flowchart of the present invention;

[0023] Figure 2 This is a flowchart illustrating the methods for the training and validation phases of a neural network. Detailed Implementation

[0024] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0025] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0026] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0027] like Figure 1 and Figure 2 As shown, this invention addresses the limitation that the parameters of the Peck formula are only applicable to certain deterministic tunnel conditions and overlying soil conditions. It proposes a method for determining the parameters of the Peck formula under different tunnel conditions and overlying soil conditions based on a neural network algorithm, thereby enabling the delineation of the settlement impact range during subway tunneling and providing basic data for the design of supporting structures.

[0028] (1) The general expression of the existing Peck formula is:

[0029]

[0030] S in the existing formula max It is a parameter related to the formation volume loss rate and the maximum ground settlement; i is the width coefficient of the settlement trough, which is related to geological conditions, tunnel depth, construction methods, etc. S max Both and i are difficult to express with a unified expression, as they vary with the conditions of the tunnel cross-section and the overlying soil layer. Therefore, the parameter S is obtained by fitting according to a specific tunnel condition and overlying soil layer condition. max The formula S and i only apply to the current tunnel and soil conditions. If the tunnel and overlying soil conditions are changed, the fitting formula will become invalid. However, for all tunnel and overlying soil conditions, due to the numerous combinations, establishing separate fitting formulas is clearly impractical. Therefore, this invention utilizes actual monitoring data under a finite combination of tunnel and overlying soil conditions to determine the Peck formula coefficients S. maxThe fit between i and the neural network model is then used for training to achieve the Peck formula coefficient S under given tunnel and overburden conditions. max The prediction of i will enable the determination of the settlement influence range and settlement distribution law of the cross section of the tunnel to be constructed, providing basic data for the delineation of the settlement influence range and the design of the support structure in the underground construction of the subway.

[0031] When obtaining cross-sectional monitoring data for a subway line and performing parameter fitting, S should first be... max Set the maximum settlement value when x = 0 to ensure that the fitting formula can correctly reflect the maximum settlement value of the subway cross section; then transform equation (1) into equation (2) to fit parameter i.

[0032]

[0033] If the cross-sectional distance x and the settlement Sx are known (as shown in Table 1), then first calculate S... max =4.41, then calculate i based on the data at each point. 2 The mean of the values ​​is used as the final estimate (points where x is 0 are not included in the calculation), i 2 =260.71. Therefore, the fitted Peck formula is:

[0034] Table 1 shows the fitting of Peck's formula based on measured data.

[0035]

[0036] (2) The Peck formula described above can calculate the lateral surface settlement range and settlement distribution using a specific subway tunnel cross-section. However, for different tunnel conditions and overlying soil conditions, its parameter S... max The values ​​of 'i' and 'i' should be different. When the Peck formula is used to predict the settlement of the cross-section of a tunnel about to be excavated, the values ​​of its parameters are difficult to determine due to the lack of measured data. Therefore, this invention combines neural networks to determine the parameters of the Peck formula for the cross-section of a tunnel about to be excavated.

[0037] First, the training set for the neural network is obtained. Parameters that significantly influence surface settlement are considered as input variables for the BP neural network: the geometric conditions of each tunnel cross-section (tunnel height, tunnel span, tunnel depth), geological conditions (elastic modulus, internal friction angle, cohesion, Poisson's ratio), and tunnel excavation speed – these nine variables are used as input variables. The Peck formula parameters S of each tunnel cross-section obtained from monitoring are then used as input variables. max and i are used as the model output. The training dataset consists of actual monitored settlement data from constructed tunnel sections.

[0038] Training and validation of the neural network model. Validation data consisted of actual settlement data monitored in the constructed tunnel section, but these were not included in the training dataset.

[0039] c. Using a trained and validated neural network model, the Peck formula parameters S for subway tunnel excavation are calculated. max The identification of i. For a tunnel section about to be excavated by cut-and-cover method, input the following nine variables: geometric conditions (tunnel height, tunnel span, tunnel depth), geological conditions (elastic modulus, internal friction angle, cohesion, Poisson's ratio), and tunnel excavation speed. This will yield the Peck formula parameters S for the tunnel cross-section to be constructed. max and i.

[0040] d will identify the Peck formula parameter S max Substituting i into formula (1) yields the Peck formula for the cross section of the tunnel to be constructed, which determines the settlement influence range and settlement distribution law of the cross section of the tunnel to be constructed, providing basic data for the delineation of the settlement influence range and the design of the support structure in the underground excavation of the subway.

[0041] The proposed method is used to predict the settlement during the construction of a tunnel using the cut-and-cover method.

[0042] (1) Monitoring data of 31 sections were collected. Only the data of cross section 1 is listed. The data of the other sections are in the same format, as shown in Table 2.

[0043] Table 2 Monitoring data of section 1

[0044]

[0045] (2) Using formula (2), the Peck formula parameter values ​​S for 31 cross sections are obtained by fitting. max And i, construct the neural network input and output datasets for each section, as shown in Table 3.

[0046] Table 3 shows the input and output datasets of the neural networks constructed based on the monitoring data for each cross-section.

[0047]

[0048] (3) A training set is constructed based on the neural network input and output data of each cross section for training and validation;

[0049] (4) Prediction of transverse surface settlement of overlying soil layer.

[0050] For a given tunnel cross-section and overlying soil parameters, as shown in Table 4, the Peck formula parameter values ​​S for this cross-section are calculated. max The values ​​of lateral surface subsidence are estimated by i and then obtained.

[0051] Table 4. Overlying soil parameters of the cross section to be estimated

[0052]

[0053] The settlement distribution curve of the overlying soil layer is as follows:

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for predicting the lateral settlement of the overlying soil layer during the cut-and-cover construction of a subway tunnel, characterized in that: The method includes the following steps: S1: The expression for Peck's formula is defined as follows: （1） S max These are parameters related to the formation volume loss rate and the maximum ground subsidence. i is the width coefficient of the settlement trough; x represents the distance of the observation point from the maximum ground settlement. When performing parameter fitting based on known subway cross-section monitoring data, S max Set as x The maximum settlement value when = 0 is determined to ensure that the fitting formula can correctly reflect the maximum settlement value of the subway cross section; Equation (1) is transformed into Equation (2) for parameterization. i The fit; （2） Known cross-sectional distance x and settlement S x ,as well as S max Substitute the data from each point into formula (2) to obtain the result. i 2 The mean as i 2 Final estimate, of which x Points with a value of 0 are not included in the calculation; based on the known cross-section... S max as well as i 2 The final estimate yields the Peck formula; repeat the above steps, perform parameter fitting based on multiple known subway cross-section monitoring data, and obtain the corresponding Peck formula; S2: Obtain the neural network training set; use the geometric conditions, geological conditions, and tunnel excavation speed of each tunnel cross-section as input variables; and use the Peck formula parameters of each tunnel cross-section obtained from monitoring. S max and i As the model output; the training dataset consists of actual settlement data monitored in constructed tunnel sections; S3: Training and validating the neural network model; S4: Using a trained and validated neural network model, perform Peck formula parameter determination for subway tunnel construction. S max and i The identification process involves inputting the geometric and geological conditions of the cross-section and the tunnel excavation speed for a tunnel section about to be excavated using the cut-and-cover method. This yields the Peck formula parameters for the cross-section of the tunnel to be constructed. S max and i ; S5: Identify the Peck formula parameters S max and i Substituting into formula (1), we obtain the Peck formula for the cross section of the tunnel to be constructed, so as to determine the settlement influence range and settlement distribution law of the cross section of the tunnel to be constructed.

2. The method for predicting the lateral settlement of the overlying soil layer during subway tunnel excavation according to claim 1, characterized in that: The geometric conditions include tunnel height, tunnel span, and tunnel depth, while the geological conditions include elastic modulus, internal friction angle, cohesion, and Poisson's ratio.

Images