Tunnel underpass existing subway line settlement control method, equipment and medium based on LSTM

Abstract

The application discloses a tunnel underpassing existing subway line settlement control method and equipment based on LSTM, and a medium, the method comprises the following steps: obtaining underpassing stratum parameters and settlement values and target lifting values of an existing line bed at a prediction time, and obtaining grouting parameters (grouting pressure, grouting amount and grouting time length) at a previous time before prediction; inputting the obtained data into a trained prediction model based on LSTM to obtain horizontal lifting grouting parameters at the prediction time; if the obtained grouting parameters are in a suitable grouting parameter solution space, the grouting parameters are taken as optimal grouting parameters at the prediction time; otherwise, a particle swarm algorithm is used to solve the optimal grouting parameters at the prediction time; the optimal grouting parameters are used for grouting at the prediction time to lift the existing subway line bed; rolling prediction and grouting are performed until the existing subway line bed is lifted to a settlement value of 0. The application can quickly obtain more reasonable lifting grouting parameters at each time, and ensures the lifting effect on the existing subway line.

Description

Technical Field

[0001] This invention belongs to the field of tunnel construction technology, and in particular relates to a settlement control method for tunnels passing under existing subway lines based on LSTM. Background Technology

[0002] With the rapid development of urban subways, complex geological conditions and the crisscrossing existing subway lines have greatly increased the difficulty of constructing new subway lines. For example, tunneling under existing subway lines or stations will affect both the construction of the new line and the operational safety of the existing subway lines. To reduce the impact of tunnel construction under existing subway lines, pre-grouting reinforcement of the surrounding strata and pipe roof / pipe curtain pre-support measures are typically implemented around the new subway tunnel, as well as pre-grouting reinforcement of the lower strata and pre-reinforcement of the internal structure around the existing subway tunnel. However, in actual projects, due to the influence of geological conditions, the cross-sectional shape and dimensions of the new tunnel and the construction methods, and the structural type and health status of the existing tunnel, the settlement of the existing subway line cannot be effectively controlled even after the above control measures are taken, resulting in settlement exceeding the limit.

[0003] For issues related to excessive settlement of existing subway lines caused by tunnel construction, grouting and lifting measures are generally adopted to restore the lost elevation of the existing subway lines. However, existing grouting and lifting measures are mainly based on the grouting parameters provided in the design, and cannot be dynamically adjusted according to geological conditions, on-site construction conditions, and the settlement control status of the existing subway lines, resulting in less than ideal lifting effects. Summary of the Invention

[0004] To quickly obtain more reasonable grouting parameters, this invention proposes a settlement control method for tunnels passing under existing subway lines based on LSTM. Based on the stratum parameters, the settlement value of the existing subway line track bed, the target uplift value, and the grouting parameters at the previous moment, the grouting parameters at the next moment are adjusted in real time. This method can ensure the safety of tunnel construction and the operation safety of the existing subway line, and has certain timeliness and economy.

[0005] To achieve the above technical objectives, the present invention adopts the following technical solution:

[0006] A settlement control method for tunnels passing under existing subway lines based on LSTM includes:

[0007] Step 1: Obtain the parameters of the strata to be crossed at the predicted time and the settlement value and target uplift value of the existing subway line track bed; obtain the horizontal uplift grouting parameters at the time before the prediction; the horizontal uplift grouting parameters include: grouting pressure, grouting volume and grouting duration;

[0008] Step 2: Input the data obtained in Step 1 into the trained LSTM-based prediction model and output the horizontal lifting grouting parameters for the predicted time.

[0009] Step 3: If the horizontal lifting grouting parameters obtained in Step 2 are within the space of suitable grouting parameter solutions, then use them as the optimal horizontal lifting grouting parameters for the prediction time; otherwise, use the particle swarm optimization algorithm and combine it with the space of suitable grouting parameter solutions to explore the optimal horizontal lifting grouting parameters, and use them as the optimal grouting parameters for the prediction time.

[0010] Step 4: At the predicted time, the optimal horizontal lifting grouting parameters are used to lift the existing subway track bed.

[0011] Step 5: Update the next prediction time and repeat steps 1-4 until the existing subway track bed is raised to a settlement value of 0.

[0012] Furthermore, the parameters of the underlying strata include: soil density, cohesion, internal friction angle, void ratio, and elastic modulus.

[0013] Furthermore, the method for determining the space of the suitable grouting parameter solution is as follows:

[0014] (1) Collect the settlement value h of the newly built tunnel arch, the stress value σ of the initial support concrete, and the horizontal lifting grouting parameters: grouting pressure P, grouting volume Q, and grouting time t;

[0015] (2) Based on the data collected in step (1), establish the relationship function between the horizontal lifting grouting parameters and the tunnel arch settlement value h and the initial support concrete stress value σ:

[0016] (P,Q,t)=F(h,σ);

[0017] (3) Determine the threshold h for tunnel arch settlement according to technical specifications. max and the threshold σ of the initial support concrete stress value max Then set the threshold h max and σ max Substituting the relational function obtained in step (2), calculate the threshold of the horizontal lifting grouting parameters: P max Q max t max Therefore, the space for suitable grouting parameters is constructed as follows: grouting pressure 0~P max Grouting volume 0~Q max Grouting time 0~t max .

[0018] Furthermore, by employing a particle swarm optimization algorithm and combining it with the solution space of suitable grouting parameters, the optimal horizontal lifting grouting parameters are explored, specifically:

[0019] (1) Use grouting pressure, grouting volume and grouting time to represent the position of the particles, initialize the particle swarm size, and use the space of the solution of the appropriate grouting parameters to limit the solution space of the particles, and assign each particle a random initial position and velocity.

[0020] (2) Based on the existing subway track bed settlement value s, the tunnel arch settlement value h, and the initial support concrete stress value σ, calculate the fitness function value f = minf(s,h,σ) of each particle in the particle swarm.

[0021] (3) Based on the current fitness function value of each particle, obtain the local optimal solution of each particle and the global optimal solution of the particle swarm;

[0022] (4) Update the velocity and position of each particle in the particle swarm, return to step (2), until the maximum number of iterations is reached;

[0023] (5) Take the global optimal solution of the particle swarm and use its corresponding grouting parameters as the optimal horizontal lifting grouting parameters.

[0024] Furthermore, the training samples and model training method of the LSTM-based prediction model are as follows:

[0025] (1) Collect the geological parameters of the underpass section based on the field exploration data, as well as the time change sequence of the existing subway track bed settlement value, target track bed uplift value, and horizontal uplift grouting parameters during the historical horizontal uplift process for the existing subway track bed settlement, and construct a training sample set.

[0026] (2) The current settlement value and target uplift value of the existing subway line track bed, as well as the soil density, cohesion, and internal friction of the strata, are used to determine the current time. The porosity, elastic modulus, and the horizontal lifting grouting parameters of the previous moment are all used as input data for the prediction model based on the LSTM long and short time series neural network; the current horizontal lifting grouting parameters are used as output data; the prediction model is trained to obtain the trained LSTM-based prediction model.

[0027] Furthermore, the settlement value of the existing subway track bed is used as the target lifting value of the track bed.

[0028] An electronic device includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, causes the processor to implement the settlement control method for tunnels passing under existing subway lines as described above.

[0029] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the settlement control method for tunnels passing under existing subway lines as described in any of the preceding claims.

[0030] Beneficial effects

[0031] 1. This invention uses the LSTM long and short time series neural network model, which is trained based on data such as the settlement value of the existing track bed, stratum parameters, horizontal lifting grouting parameters and the lifting value of the existing subway track bed in the previous period. It can quickly obtain more reasonable lifting grouting parameters for the next moment, ensuring the lifting effect of the existing subway line.

[0032] 2. This invention, while fully considering the impact of grouting and lifting measures on the support structure of the new tunnel, effectively controls the settlement of the existing subway line caused by the underpass construction, ensuring the safety of the new tunnel underpass construction and the operational safety of the existing subway line. Attached Figure Description

[0033] Figure 1 This is a flowchart of the method of the present invention;

[0034] Figure 2 A frontal view of the horizontal lifting grouting construction for the tunnel passing under the existing subway line;

[0035] Figure 3 A side view of the horizontal lifting and grouting construction for the tunnel passing under the existing subway line. Detailed Implementation

[0036] The embodiments of the present invention will be described in detail below. These embodiments are based on the technical solutions of the present invention and provide detailed implementation methods and specific operation processes to further explain the technical solutions of the present invention.

[0037] See Figures 1-3 A settlement control method for tunnels passing under existing subway lines based on LSTM includes the following steps:

[0038] S1: Based on the field exploration data, collect the stratum parameters (such as density, cohesion, internal friction angle, void ratio, and elastic modulus) of the section under which the new tunnel 3 passes, the settlement monitoring values ​​of the track bed 2 of the existing metro line 1, the grouting parameters (such as grouting pressure, grouting volume, and grouting time) of the horizontal lifting grouting measure 4 at the pit end 5, and the target lifting value of the track bed 2, and construct a training sample set.

[0039] S2: Construct an LSTM long and short time series neural network model, which incorporates the current settlement monitoring value (s) of the existing subway track bed 2, the target uplift value (h) of track bed 2, and the soil density (ρ), cohesion (c), and internal friction angle of the stratum. The porosity (e), elastic modulus (E), and grouting parameters from the previous moment: grouting pressure (P), grouting volume (Q), and grouting duration (t) are used as input data for the LSTM long and short time series neural network prediction model. The grouting parameters for the current moment of horizontal lifting grouting 4: grouting pressure (P), grouting volume (Q), and grouting duration (t) are used as output data for the prediction model. The model is trained to obtain the relationship between the parameters, thus obtaining the trained LSTM-based horizontal lifting grouting parameter prediction model.

[0040] S3: Obtain the parameters of the underlying strata and the settlement and target uplift values ​​of the existing subway line track bed at the predicted time. Obtain the horizontal uplift grouting parameters at the time before the prediction, wherein the initial horizontal uplift grouting parameters are selected from the grouting parameters at the moment when the shield tunnel is about to enter the existing line. In this embodiment, the settlement value of the existing subway line track bed is taken as the target uplift value of the track bed.

[0041] S4: Input the data obtained in S3 into the trained LSTM-based prediction model and output the horizontal lifting grouting parameters at the predicted time.

[0042] S5: If the horizontal lifting grouting parameters obtained in S4 are within the space of suitable grouting parameter solutions, then use them as the optimal horizontal lifting grouting parameters for the prediction time; otherwise, use the particle swarm optimization algorithm and combine it with the space of suitable grouting parameter solutions to explore the optimal horizontal lifting grouting parameters, and use them as the optimal grouting parameters for the prediction time.

[0043] The method for determining the space of the suitable grouting parameter solution is as follows:

[0044] (1) Collect the settlement value h of the newly built tunnel arch, the stress value σ of the initial support concrete, and the horizontal lifting grouting parameters: grouting pressure P, grouting volume Q, and grouting time t;

[0045] (2) Based on the data collected in step (1), establish the relationship function between the horizontal lifting grouting parameters and the tunnel arch settlement value h and the initial support concrete stress value σ:

[0046] (P,Q,t)=F(h,σ);

[0047] (3) Determine the threshold h for tunnel arch settlement according to the "Technical Specification for Safety Protection of Urban Rail Transit Structures" (CJJ T 202-2013). max The threshold value σ of the initial support concrete stress is determined according to the "Code for Design of Concrete Structures" (GB 50010-2010). max Then set the threshold h max and σ max Substituting the relational function obtained in step (2), calculate the threshold of the horizontal lifting grouting parameters: P max Q maxt max Therefore, the space for suitable grouting parameters is constructed as follows: grouting pressure 0~P max Grouting volume 0~Q max Grouting time 0~t max .

[0048] In addition, the optimal horizontal lifting grouting parameters are explored by employing a particle swarm optimization algorithm combined with the solution space of appropriate grouting parameters, specifically:

[0049] (1) The velocity of the particles is represented by grouting pressure, grouting volume and grouting time. The number of particles is initialized to 30. The solution space of the particles is restricted by the solution space of the appropriate grouting parameters. Each particle is assigned a random initial position and velocity.

[0050] (2) Based on the existing subway track bed settlement value s, the tunnel arch settlement value (h), and the initial support concrete stress value σ, calculate the fitness function value f = minf(s,h,σ) of each particle in the particle swarm.

[0051] (3) Based on the current fitness function value of each particle, obtain the local optimal solution of each particle and the global optimal solution of the particle swarm;

[0052] (4) Update the velocity and position of each particle in the particle swarm, return to step (2), until the maximum number of iterations is reached;

[0053] (5) Take the global optimal solution of the particle swarm and use its corresponding grouting parameters as the optimal horizontal lifting grouting parameters.

[0054] S6: At the predicted time, the optimal horizontal lifting grouting parameters are used to lift the existing subway track bed.

[0055] S7: Update the next prediction time and repeat steps S3-S6 to adjust the grouting parameters in real time until the existing subway track bed is raised to a settlement value of 0, ensuring the grouting lifting effect and effectively controlling the elevation loss of the existing subway track bed and the structural safety of the new tunnel.

[0056] The above embodiments are preferred embodiments of this application. Those skilled in the art can make various changes or improvements based on them. Without departing from the overall concept of this application, these changes or improvements should fall within the scope of protection claimed in this application.

Claims

1. A settlement control method for tunnels passing under existing subway lines based on LSTM, characterized in that, include: Step 1: Obtain the geological parameters of the strata to be crossed at the predicted time, as well as the settlement and target uplift values ​​of the existing subway track bed. Also obtain the horizontal uplift grouting parameters for the time preceding the prediction. These horizontal uplift grouting parameters include: grouting pressure. Grouting volume and grouting duration ; Step 2: Input the data obtained in Step 1 into the trained LSTM-based prediction model and output the horizontal lifting grouting parameters for the predicted time. Step 3: If the horizontal lifting grouting parameters obtained in Step 2 are within the space of suitable grouting parameter solutions, then use them as the optimal horizontal lifting grouting parameters for the prediction time; otherwise, use the particle swarm optimization algorithm and combine it with the space of suitable grouting parameter solutions to explore the optimal horizontal lifting grouting parameters, and use them as the optimal grouting parameters for the prediction time. Specifically, the optimal horizontal lifting grouting parameters are explored by employing a particle swarm optimization algorithm combined with the solution space of appropriate grouting parameters. (1) Use grouting pressure, grouting volume and grouting time to represent the position of the particles, initialize the particle swarm size, and use the space of the solution of the appropriate grouting parameters to limit the solution space of the particles, and assign each particle a random initial position and velocity. (2) Based on the existing subway track bed settlement value s, tunnel arch settlement value h, and initial support concrete stress value Calculate the fitness function value of each particle in the particle swarm. ; (3) Based on the current fitness function value of each particle, obtain the local optimal solution of each particle and the global optimal solution of the particle swarm; (4) Update the velocity and position of each particle in the particle swarm, and return to step (2) until the maximum number of iterations is reached; (5) Take the global optimal solution of the particle swarm and use its corresponding grouting parameters as the optimal horizontal lifting grouting parameters; The method for determining the space of suitable grouting parameters is as follows: The threshold value of tunnel arch settlement is determined according to technical specifications. Threshold for initial support concrete stress value Then set the threshold and Substitute into relational function Calculate the threshold values ​​for horizontal lift grouting parameters: , , Therefore, the space for suitable grouting parameter solutions is constructed as follows: grouting pressure 0~ Grouting volume 0~ Grouting time 0~ ; Step 4: At the predicted time, the optimal horizontal lifting grouting parameters are used to lift the existing subway track bed. Step 5: Update the next prediction time and repeat steps 1-4 until the existing subway track bed is raised to a settlement value of 0.

2. The method according to claim 1, characterized in that, The parameters of the underlying strata include: soil density, cohesion, internal friction angle, void ratio, and elastic modulus.

3. The method according to claim 1, characterized in that, The training samples and model training method of the LSTM-based prediction model are as follows: (1) Collect the geological parameters of the underpass section based on the field exploration data, as well as the time change sequence of the existing subway line track bed settlement value, track bed target uplift value, and horizontal uplift grouting parameters during the historical horizontal uplift process for the existing subway line track bed settlement, and construct a training sample set; (2) The current settlement value and target uplift value of the existing subway line track bed, soil density, cohesion and internal friction angle of the stratum. The porosity, elastic modulus, and the horizontal lifting grouting parameters of the previous moment are all used as input data for the prediction model based on LSTM long and short time series neural network; the current horizontal lifting grouting parameters are used as output data; the prediction model is trained to obtain a trained LSTM-based prediction model.

4. The method according to claim 1, characterized in that, The settlement value of the existing subway track bed is used as the target lifting value of the track bed.

5. An electronic device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the computer program is executed by the processor, the processor causes the processor to implement the method as described in any one of claims 1 to 4.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 4.

Images