A prediction method for ground settlement during the excavation of overlapping shield tunnels

Abstract

The present invention discloses a method for predicting the ground settlement during the excavation of overlapping shield tunnels. The method includes: obtaining tunnel and stratum data, inputting the tunnel and stratum data into a prediction model, and obtaining the influence parameters of the existing tunnel and the influence parameters of the new tunnel output by the prediction model; establishing a ground settlement curve of the overlapping shield tunnel according to the influence parameters of the existing tunnel and the influence parameters of the new tunnel; predicting the ground settlement during the excavation of the overlapping shield tunnel according to the ground settlement curve of the overlapping shield tunnel. Based on knowledge-driven, the present invention takes into account that the ground settlement curve of the overlapping tunnel is affected by both the existing tunnel and the new tunnel, and clarifies that the influence parameters of the existing tunnel and the influence parameters of the new tunnel need to be obtained to establish the ground settlement curve of the overlapping tunnel. Then, based on data-driven, the influence parameters of the existing tunnel and the influence parameters of the new tunnel are obtained through a neural network, so as to realize the prediction of the ground settlement during the excavation of the overlapping shield tunnel.

Description

Technical Field

[0001] The present invention relates to the technical field of underground structure prediction, and particularly to a method for predicting ground settlement during the excavation of overlapping shield tunnels. Background Art

[0002] It is generally believed in the prior art that the form of the ground settlement curve caused by the excavation of a single shield tunnel is a Gaussian distribution curve.

[0003] In the ground settlement curve in the form of a Gaussian distribution curve, to obtain the ground settlement curve caused by the excavation of a single tunnel, only the maximum ground settlement value at the tunnel axis needs to be obtained. For this purpose, the prior art usually adopts two methods for processing. One is the knowledge-driven method, and the other is the data-driven method.

[0004] Through the knowledge-driven method, a formula for calculating the maximum ground settlement value is obtained through multiple experiments, and the maximum ground settlement value is calculated through this formula.

[0005] On the other hand, in the prior art, the maximum ground settlement is predicted by using a neural network through data driving. For example, a hybrid prediction model of an artificial neural network optimized by particle swarm optimization is used to predict the maximum ground settlement, or three neural network models with the parameters of earth pressure balance shield and geological conditions as inputs and the maximum settlement as outputs are established.

[0006] However, the current methods using neural network prediction are only used to predict the maximum settlement value. This method can only be used for the prediction of single-line tunnels. In densely populated and prosperous areas, with the continuous development of the underground space, overlapping tunnels, with their good ability to adapt to the construction conditions of restricted spaces, are increasing in the development of the underground space in prosperous urban areas. Compared with single-line tunnels, the settlement caused by the excavation of overlapping tunnels is affected by both the existing tunnel and the newly built tunnel. Therefore, the ground settlement curve of overlapping shield tunnels cannot be represented by a simple Gaussian distribution curve. It has characteristics such as an uncertain maximum settlement position and multiple peaks. Coupled with the fact that the traditional neural network method can only better predict the maximum settlement value and has poor predictability for the settlement curve distribution, the prediction of the ground settlement curve caused by the excavation of overlapping tunnels faces challenges, resulting in difficulties in predicting the ground settlement of overlapping shield tunnels. Summary of the Invention

[0007] The purpose of the present invention is to provide a method for predicting ground settlement during the excavation of overlapping shield tunnels, aiming to solve the problem of difficult prediction of ground settlement of overlapping shield tunnels in the prior art.

[0008] The technical solution adopted by the present invention to solve the technical problem is as follows:

[0009] The present invention provides a method for predicting the ground settlement during the excavation of overlapping shield tunnels. The method for predicting the ground settlement during the excavation of overlapping shield tunnels includes:

[0010] Obtain tunnel and stratum data, input the tunnel and stratum data into a prediction model, and obtain the influence parameters of the existing tunnel and the influence parameters of the new tunnel output by the prediction model;

[0011] Establish a ground settlement curve for the overlapping shield tunnels based on the influence parameters of the existing tunnel and the influence parameters of the new tunnel;

[0012] Predict the ground settlement during the excavation of the overlapping shield tunnels according to the ground settlement curve of the overlapping shield tunnels.

[0013] Further, the tunnel and stratum data include the tunnel diameter, the buried depth of the existing tunnel, the horizontal distance between the existing tunnel and the new tunnel, the vertical distance between the existing tunnel and the new tunnel, the ground loss rate of the existing tunnel, and the ground loss rate of the new tunnel as the tunnel and stratum data.

[0014] Further, the obtaining of the tunnel and stratum data specifically includes:

[0015] Obtain the tunnel diameter, the buried depth of the existing tunnel, the horizontal distance between the existing tunnel and the new tunnel, and the vertical distance between the existing tunnel and the new tunnel according to the tunnel construction drawings;

[0016] Obtain the ground loss rate of the existing tunnel and the ground loss rate of the new tunnel according to the type of shield machine used in the tunnel construction.

[0017] Further, the inputting of the tunnel and stratum data into the prediction model to obtain the influence parameters of the existing tunnel and the influence parameters of the new tunnel output by the prediction model specifically includes:

[0018] Establish a prediction model, where the prediction model includes an input layer, a pattern layer, a summation layer, and an output layer;

[0019] Input the tunnel and stratum data into the prediction model, and the input layer transmits the tunnel and stratum data to the pattern layer;

[0020] The pattern layer processes the tunnel and stratum data using a Gaussian function and then outputs;

[0021] The summation layer performs arithmetic mean and multiple weighted means on the outputs of the pattern layer respectively;

[0022] The output layer outputs the influence parameters of the existing tunnel and the influence parameters of the new tunnel according to the arithmetic mean and the weighted mean.

[0023] Further, the establishment of the prediction model specifically includes:

[0024] Obtain a training data set and establish an initial model;

[0025] Train the initial model with the training data set to obtain the prediction model.

[0026] Further, the obtaining of the training data set specifically includes:

[0027] Obtain training data of multiple tunnels and strata data;

[0028] Conduct finite element simulations on the training data of multiple tunnels and strata data, and obtain the existing tunnel influence parameter training data and tunnel influence parameter training data corresponding to each tunnel and strata data training data;

[0029] Form multiple training samples with each of the tunnel and strata data training data, and the corresponding existing tunnel influence parameter training data and the corresponding tunnel influence parameter training data;

[0030] Use the set of multiple training samples as the training data set.

[0031] Further, the conducting of finite element simulations on the training data of multiple tunnels and strata data to obtain the existing tunnel influence parameter training data and tunnel influence parameter training data corresponding to each tunnel and strata data training data specifically includes:

[0032] Conduct finite element simulations on the training data of multiple tunnels and strata data to obtain the corresponding ground settlement data set;

[0033] Fit the ground settlement data set to obtain the corresponding existing tunnel influence parameter training data and tunnel influence parameter training data.

[0034] Further, the training of the initial model with the training data set to obtain the prediction model specifically includes:

[0035] Divide the training data set into a training set and a validation set according to a set division ratio;

[0036] Calculate the set evaluation parameters of each training sample in the training set, and move the training samples with the set evaluation parameters lower than the set threshold to the validation set;

[0037] Train the initial model according to the training set and the validation set to obtain the prediction model.

[0038] Further, the calculation of the set evaluation parameters of each training sample in the training set specifically includes:

[0039] Calculate the correlation coefficient R of the training sample based on the surface settlement data set corresponding to the training sample, the influence parameter training data of the existing tunnel, and the influence parameter training data of the tunnel 2 ;

[0040] Use the correlation coefficient R 2 as the set evaluation parameter.

[0041] Furthermore, the influence parameters of the existing tunnel include the peak influence of the existing tunnel settlement , the influence deviation parameter of the existing tunnel and the width coefficient of the surface settlement trough of the existing tunnel , and the influence parameters of the new tunnel include the peak influence of the new tunnel settlement , the influence deviation parameter of the new tunnel and the width coefficient of the surface settlement trough of the new tunnel ;

[0042] Establishing the surface settlement curve of the overlapping shield tunnel according to the influence parameters of the existing tunnel and the influence parameters of the new tunnel specifically includes:

[0043] According to the peak influence of the existing tunnel settlement , the influence deviation parameter of the existing tunnel , the width coefficient of the surface settlement trough of the existing tunnel , the peak influence of the new tunnel settlement , the influence deviation parameter of the new tunnel and the width coefficient of the surface settlement trough of the new tunnel Establish the surface settlement curve of the overlapping shield tunnel:

[0044] ;

[0045] where represents the surface settlement value at.

[0046] The present invention has the following effects when adopting the above technical solutions:

[0047] The present invention, driven by knowledge, takes into account that the overlapping tunnel is affected by both the existing tunnel and the new tunnel. It is clear that obtaining the influence parameters of the existing tunnel and the new tunnel is required to establish the surface settlement curve of the overlapping tunnel. Then, driven by data, the influence parameters of the existing tunnel and the new tunnel are obtained through a neural network. Thus, based on the dual drive of data and knowledge, the ground settlement curve of the overlapping tunnel is established to realize the prediction of the surface settlement during the excavation of the overlapping shield tunnel. Description of the Drawings

[0048] Figure 1 It is a flowchart of the steps of a method for predicting ground settlement during the excavation of overlapping shield tunnels in a preferred embodiment of the present invention;

[0049] Figure 2 It is a structural flowchart of a method for predicting ground settlement during the excavation of overlapping shield tunnels in a preferred embodiment of the present invention;

[0050] Figure 3 It is a schematic diagram of the ground settlement curve of overlapping shield tunnels;

[0051] Figure 4 It is a schematic diagram of the structure of the generalized neural network model in a preferred embodiment of the present invention. Detailed implementation manners

[0052] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to 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 the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0053] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, device, product or terminal that includes a series of steps or units is not limited to the listed steps or units, but may optionally further include steps or units not listed, or may optionally further include other steps or units inherent to these processes, methods, products or terminals.

[0054] Referring to "embodiment" in this context means that a specific feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the present invention. The phrase appears in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein may be combined with other embodiments.

[0055] Embodiment 1

[0056] Please refer to Figure 1 , Embodiment 1 of the present application is a method for predicting ground settlement during the excavation of overlapping shield tunnels, wherein, Figure 1The described method can be applied to a device for predicting ground settlement during the excavation of overlapping shield tunnels. Among them, the device may include a server, where the server includes a local server or a cloud server, which is not limited in the embodiments of the present invention.

[0057] Please refer to Figure 1 and Figure 2 as shown, Figure 1 and Figure 2 The method shown includes the steps:

[0058] S1. Obtain tunnel and stratum data, input the tunnel and stratum data into the prediction model, and obtain the influence parameters of the existing tunnel and the influence parameters of the new tunnel output by the prediction model.

[0059] S2. Establish a ground settlement curve for overlapping shield tunnels based on the influence parameters of the existing tunnel and the influence parameters of the new tunnel.

[0060] Specifically, please refer to Figure 3 , the ground settlement curve of overlapping tunnels is different from that of single tunnels. For single tunnels, the position where the maximum settlement occurs is near the tunnel centerline. However, for overlapping tunnels, due to the redistribution of soil stress caused by stratum excavation, the position where the maximum settlement occurs is not necessarily at the tunnel axis. In most cases, it is between the two tunnel axes, and the specific position is difficult to predict.

[0061] In addition, the shape of the ground settlement curve of overlapping shield tunnels cannot be directly represented by a Gaussian curve like that of single tunnels. Due to the different distances between tunnels, the superposition and mutual influence of the displacement field and stress field result in different shapes of the ground settlement curve, which may be single-peak, double-peak, multi-peak, the curve may be approximately symmetric or asymmetric, which means that key indicators such as the maximum differential settlement are difficult to predict.

[0062] The above two reasons first make it difficult to predict the maximum settlement of the settlement curve of overlapping tunnels, and also make it difficult to deduce the complete settlement curve even if the maximum settlement of the settlement curve is successfully predicted.

[0063] In the prior art, the theoretical solutions for overlapping tunnels have been gradually proposed in recent years. For example: One prior art proposed the ground settlement curve caused by the excavation of shallow overlapping tunnels in soft soil, but its correction is only applicable to shallow soft soil strata. Another prior art corrected the application of the ground settlement curve of single tunnels in the upper and lower overlapping tunnels by calculating the stratum loss rate with the burial depth as the weight, but its correction formula is only applicable to the working conditions of upper and lower overlapping tunnels. Another prior art proposed an expression for the ground settlement curve under any arrangement of double-track overlapping tunnels, but its form is complex and its application is limited. In knowledge-driven prediction methods, theoretical formulas are mostly not universal and can only be applied to specific situations.

[0064] Therefore, for the above reasons, in this embodiment, based on knowledge-driven, considering that the overlapping tunnels are affected by both the existing tunnel and the newly built tunnel, it is clear that obtaining the influence parameters of the existing tunnel and the newly built tunnel is required to establish the surface settlement curve of the overlapping tunnels. Then, based on data-driven, the influence parameters of the existing tunnel and the newly built tunnel are obtained through a neural network, so as to be driven by both data and knowledge for subsequent establishment of the ground settlement curve of the overlapping tunnels.

[0065] Specifically, in this embodiment, first, according to the characteristics of the surface settlement curve distribution caused by the excavation of the overlapping tunnels, assuming that there are two overlapping tunnels, the one constructed first is called the existing tunnel, and the one constructed later is called the newly built tunnel. Due to the coupling effect among the existing tunnel - formation - newly built tunnel, the form of the surface settlement curve shows different properties from that of a single-line tunnel, including but not limited to: the measured value of the ground displacement caused by the overlapping excavation of the two tunnels up and down is often smaller than the theoretical superposition value caused by the excavation of a single tunnel up and down; compared with the excavation of a single-line tunnel, the newly built tunnel will change the stress state of the formation around the existing tunnel, thus affecting the size of the settlement trough and the maximum surface settlement; the influence range, surface settlement and horizontal displacement caused by the construction of the overlapping tunnels are all smaller than those of the parallel closely spaced tunnels, but the local curvature and slope of the settlement trough are much larger than those of the horizontally spaced tunnels.

[0066] Based on the above knowledge-driven, although the coupling effect among the existing tunnel - formation - newly built tunnel is very complex and there is no unified theory and widely recognized explanation, an approximately fitting empirical formula can still be proposed according to past research and experience. In this embodiment, an expression of the surface settlement curve of the overlapping shield tunnels containing 6 parameters with clear physical meanings is proposed:

[0067] ;

[0068] Among them, represents the peak value of the settlement influence of the existing tunnel, represents the surface settlement value at, represents the influence deviation parameter of the existing tunnel, which is a parameter that causes the settlement curve to shift due to the horizontal distance between the two tunnels in the settlement trough, is the width coefficient of the surface settlement trough of the existing tunnel, represents the peak value of the settlement influence of the newly built tunnel, represents the influence deviation parameter of the newly built tunnel, represents the width coefficient of the surface settlement trough of the newly built tunnel.

[0069] It can be seen that, aiming at the problems existing in the prediction of the surface settlement curve by the existing neural network, namely, the peak value is difficult to predict and there is a lack of interpretability, the expression of the surface settlement curve of the overlapping shield tunnel proposed in this embodiment can represent the general form of the surface settlement, and the parameters therein have relatively clear meanings, so as to enhance the interpretability of the prediction of the subsequent neural network prediction model, and the parameters are appropriate to simplify the model.

[0070] In this embodiment, a neural network is used as a prediction model to predict the peak value of the settlement influence of the existing tunnel , the influence deviation parameter of the existing tunnel , the width coefficient of the surface settlement trough of the existing tunnel , the peak value of the settlement influence of the new tunnel , the influence deviation parameter of the new tunnel and the width coefficient of the surface settlement trough of the new tunnel , and substitute these parameters into the proposed surface settlement curve expression, so as to predict the surface settlement curve caused by the excavation of the overlapping tunnel.

[0071] In addition, in this embodiment, based on knowledge-driven, appropriate parameters are selected from existing studies for subsequent analysis by the neural network. In this embodiment, the specifically selected parameters include the tunnel diameter (the diameters of the existing tunnel and the new tunnel are the same), the buried depth of the existing tunnel , the ground loss rate of the existing tunnel , the ground loss rate of the new tunnel , the horizontal distance between the existing tunnel and the new tunnel , the vertical distance between the existing tunnel and the new tunnel .

[0072] Among them, when predicting the tunnel settlement, the tunnel diameter , the buried depth of the existing tunnel , the horizontal distance between the existing tunnel and the new tunnel , the vertical distance between the existing tunnel and the new tunnel are determined through the tunnel construction drawings, and the ground loss rate of the existing tunnel and the ground loss rate of the new tunnel are determined by the type of shield machine during tunnel construction.

[0073] Since the surface settlement curve of the above-mentioned overlapping shield tunnels, although simple in form, is difficult to determine parameters through mechanical principles, in this embodiment, a neural network is used as a prediction model. By training the neural network, the prediction of the parameters of this expression is achieved. Substituting the predicted parameters back into the surface settlement expression, the settlement at any position on the surface is output, overcoming the shortcoming that the previous neural network could only predict the maximum settlement, and enabling reasonable prediction of the issues that overlapping tunnels are more concerned about compared to single-line tunnels: the position of the maximum surface settlement, the maximum non-uniform settlement, etc.

[0074] During training, first obtain the training data set. After obtaining the training data set, it is necessary to divide the training set into a training set and a test set. Due to different curve fitting effects, when dividing the test set and the training set, a division method different from the traditional method should be adopted: initially select approximately 80% of the data as the training set and 20% of the data as the test set. Then calculate the correlation coefficient R of all the surface settlement curves in the training set. 2 For the R of the expression of the surface settlement curve, 2 Those with R < 0.8 are excluded from the training set and changed to the test set, and then training is carried out according to the training set and the test set.

[0075] Specifically, please refer to Figure 4 In this embodiment, the prediction model is specifically a generalized regression neural network model, and the generalized regression neural network model includes an input layer, a pattern layer, a summation layer, and an output layer.

[0076] Among them, the number of neurons in the input layer is the same as the number of independent variables, which plays the role of transferring variables.

[0077] The number of neurons in the pattern layer is the same as the number of independent variables. The output formula of the th neuron in the pattern layer is:

[0078] ;

[0079] Among them, represents the input, and the input includes the tunnel diameter (the diameters of the existing tunnel and the new tunnel are the same), the buried depth of the existing tunnel, the ground loss rate of the existing tunnel, the ground loss rate of the new tunnel, the horizontal distance between the existing tunnel and the new tunnel, the vertical distance between the existing tunnel and the new tunnel, represents the training sample of the th neuron in the pattern layer, represents the output of the th neuron in the pattern layer, represents the smoothing factor.

[0080] The summation layer includes the first - type summation neurons and the second - type summation neurons. The number of the first - type summation neurons is 1, and it calculates the average value of the outputs of the neurons in each pattern layer:

[0081] ;

[0082] where, represents the output of the first - type summation neurons, represents the number of neurons in the pattern layer.

[0083] The number of the second - type summation neurons is the same as the number of neurons in the output layer. The second - type summation neurons calculate the weighted average value of the pattern layer:

[0084] ;

[0085] where, is the output of the th second - type summation neuron, is the weighted value assigned by the th second - type summation neuron to the output of the th neuron in the pattern layer.

[0086] The number of neurons in the output layer is the same as the number of output parameter values. Each output neuron corresponds to an output parameter, and its output formula is:

[0087] ;

[0088] where, is the output of the output - layer neuron corresponding to the th second - type summation neuron. In this embodiment, has a value of 6. The outputs of the 6 output - layer neurons respectively correspond to the peak value of the influence of the existing tunnel settlement , the deviation parameter of the influence of the existing tunnel , the width coefficient of the ground settlement trough of the existing tunnel , the peak value of the influence of the new - built tunnel settlement , the deviation parameter of the influence of the new - built tunnel and the width coefficient of the ground settlement trough of the new - built tunnel these 6 parameters.

[0089] In summary, in the present invention, considering that the overlapping tunnel is affected by both the existing tunnel and the newly built tunnel based on knowledge-driven, it is clear that the influence parameters of the existing tunnel and the newly built tunnel need to be obtained to establish the ground settlement curve of the overlapping tunnel. Then, based on data-driven, the influence parameters of the existing tunnel and the newly built tunnel are obtained through a neural network, so as to establish the ground settlement curve of the overlapping tunnel based on the dual drive of data and knowledge and realize the prediction of the ground settlement during the excavation of the overlapping shield tunnel.

[0090] It should be noted that in this article, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or terminal including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or terminal. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of another identical element in the process, method, article or terminal including that element.

[0091] Certainly, those of ordinary skill in the art can understand that all or part of the processes in the above method embodiments can be completed by instructing relevant hardware (such as a processor, a controller, etc.) through a computer program. The program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the above method embodiments. The storage medium can be a memory, a magnetic disk, an optical disk, etc.

[0092] It should be understood that the application of the present invention is not limited to the above examples. For those of ordinary skill in the art, improvements or transformations can be made according to the above description, and all such improvements and transformations should fall within the protection scope of the appended claims of the present invention.

Claims

1. A method for predicting surface settlement during excavation of overlapping shield tunnels, characterized in that: The method for predicting surface settlement during excavation of overlapping shield tunnels comprises: Acquire tunnel and stratum data, input the tunnel and stratum data into a prediction model, and obtain influencing parameters of existing tunnels and influencing parameters of newly built tunnels output by the prediction model; Establishing a surface settlement curve of a cross-overlapping shield tunnel according to the influencing parameters of the existing tunnel and the influencing parameters of the newly built tunnel; Predicting the surface settlement of the cross-overlap shield tunnel excavation according to the cross-overlap shield tunnel surface settlement curve; The existing tunnel influencing parameters include the existing tunnel settlement influencing peak value , Existing tunnel influence deviation parameters and the width coefficient of the existing tunnel surface settlement trough The new tunnel impact parameters include the new tunnel settlement impact peak value , New tunnel affects deviation parameters and the width coefficient of the surface settlement trough of the new tunnel ; The establishing of the surface settlement curve of the cross-overlapping shield tunnel according to the influencing parameters of the existing tunnel and the influencing parameters of the newly-built tunnel specifically includes: According to the existing tunnel settlement impact peak , the existing tunnel influence deviation parameter , the width coefficient of the surface settlement trough of the existing tunnel , the peak value of the settlement impact of the new tunnel , the new tunnel affects the deviation parameters and the width coefficient of the surface settlement trough of the new tunnel Establishing surface settlement curves for cross-overlapping shield tunnels: ； in, express The surface settlement value at .

2. The method for predicting surface settlement during excavation of overlapping shield tunnels according to claim 1, characterized in that: The tunnel and stratum data include tunnel diameter, burial depth of existing tunnel, horizontal distance between existing tunnel and new tunnel, vertical distance between existing tunnel and new tunnel, stratum loss rate of existing tunnel and stratum loss rate of new tunnel.

3. The method for predicting surface settlement during excavation of overlapping shield tunnels according to claim 2, characterized in that: The obtaining of tunnel and stratum data specifically includes: According to the tunnel construction drawings, the tunnel diameter, the buried depth of the existing tunnel, the horizontal distance between the existing tunnel and the newly built tunnel, and the vertical distance between the existing tunnel and the newly built tunnel are obtained; The stratum loss rate of the existing tunnel and the stratum loss rate of the newly built tunnel are obtained according to the type of shield machine used in tunnel construction.

4. The method for predicting surface settlement during excavation of overlapping shield tunnels according to claim 2, characterized in that: The step of inputting the tunnel and stratum data into the prediction model to obtain the influencing parameters of the existing tunnel and the influencing parameters of the newly built tunnel output by the prediction model specifically includes: Establishing a prediction model, wherein the prediction model includes an input layer, a pattern layer, a summation layer and an output layer; The tunnel and stratum data are input into the prediction model, and the input layer transmits the tunnel and stratum data to the model layer; The model layer processes the tunnel and stratum data using a Gaussian function and then outputs the processed data; The summation layer performs arithmetic mean and multiple weighted mean on the output of the pattern layer respectively; The output layer outputs the existing tunnel influencing parameters and the newly-built tunnel influencing parameters according to the arithmetic mean and the weighted mean.

5. The method for predicting surface settlement during excavation of overlapping shield tunnels according to claim 4, characterized in that: The establishment of the prediction model specifically includes: Obtain training data set and build initial model; The initial model is trained with the training data set to obtain the prediction model.

6. A method for predicting surface settlement during excavation of overlapping shield tunnels according to claim 5, characterized in that: The obtaining of the training data set specifically includes: Obtain multiple tunnel and formation data training data; Performing finite element simulation on the plurality of tunnel and stratum data training data to obtain existing tunnel influencing parameter training data and tunnel influencing parameter training data corresponding to each tunnel and stratum data training data; The training data of each tunnel and stratum data, the corresponding training data of the existing tunnel influencing parameters and the corresponding training data of the tunnel influencing parameters constitute a plurality of training samples; A set of multiple training samples is used as the training data set.

7. A method for predicting surface settlement during excavation of overlapping shield tunnels according to claim 6, characterized in that: The performing of finite element simulation on the plurality of tunnel and stratum data training data to obtain existing tunnel influencing parameter training data and tunnel influencing parameter training data corresponding to each tunnel and stratum data training data specifically includes: Performing finite element simulation on a plurality of the tunnel and stratum data training data to obtain corresponding surface settlement data sets; The surface settlement data set is fitted to obtain corresponding existing tunnel influencing parameter training data and tunnel influencing parameter training data.

8. The method for predicting surface settlement during excavation of overlapping shield tunnels according to claim 7, characterized in that: The step of training the initial model with the training data set to obtain the prediction model specifically includes: Dividing the training data set into a training set and a validation set according to a set division ratio; Calculating a set evaluation parameter for each of the training samples in the training set, and moving the training samples whose set evaluation parameters are lower than a set threshold to a validation set; The initial model is trained according to the training set and the validation set to obtain the prediction model.

9. A method for predicting surface settlement during excavation of overlapping shield tunnels according to claim 8, characterized in that: The calculating of the set evaluation parameters of each training sample in the training set specifically includes: Calculate the correlation coefficient R2 of the training sample according to the surface settlement data set corresponding to the training sample, the existing tunnel influencing parameter training data and the tunnel influencing parameter training data; The correlation coefficient R2 is used as the setting evaluation parameter.

Images