A method, system, terminal and storage medium for predicting deformation of a tunnel adjacent to a tunnel

Abstract

This invention relates to the field of tunnel engineering and discloses a method, system, terminal, and storage medium for predicting the deformation of adjacent tunnels. The method includes: acquiring the soil properties and basic model of the tunnel construction site; setting the parameters of the basic model according to the soil properties to obtain an initial prediction model; calibrating the initial prediction model using an optimization algorithm to obtain a standard prediction model; predicting the tunnel deformation using the standard prediction model to obtain a predicted deformation; calculating the support axial force value based on the predicted deformation using the standard prediction model to obtain a target support axial force value; and obtaining construction instructions based on the target support axial force value and applying them to tunnel construction. This invention predicts tunnel deformation using a standard prediction model, calculates the target support axial force value based on the predicted deformation, and applies the target support axial force value to tunnel construction, enabling timely monitoring of whether tunnel deformation exceeds limits and allowing for remedial measures.

Description

Technical Field

[0001] This invention relates to the field of tunnel engineering, and more particularly to a method, system, terminal, and computer-readable storage medium for predicting deformation of adjacent tunnels. Background Technology

[0002] When deep foundation pits are excavated in densely populated urban areas, safety issues often arise with adjacent existing tunnels. The unloading and soil displacement caused by excavation can easily lead to excessive settlement, heave, or elliptic deformation of the tunnel, which may threaten the structural safety and operation of the tunnel in severe cases. Therefore, it is necessary to monitor the deformation of adjacent tunnels during construction.

[0003] Existing methods for predicting the deformation of adjacent tunnels often involve setting up monitoring points during construction and making a one-time prediction using numerical methods such as the finite element method. However, this method is limited by the accuracy of the initial soil and rock parameters and can only passively observe whether the deformation exceeds the limit during construction, making it impossible to monitor whether the tunnel deformation exceeds the limit in a timely manner and take remedial measures.

[0004] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0005] The main objective of this invention is to provide a method, system, terminal, and computer-readable storage medium for predicting the deformation of adjacent tunnels. This invention aims to solve the problem that existing methods for predicting the deformation of adjacent tunnels are limited by the accuracy of initial soil and rock parameters, and can only passively observe whether the deformation exceeds the limit during construction, resulting in the inability to monitor the tunnel deformation in a timely manner and take remedial measures.

[0006] To achieve the above objectives, the present invention provides a method for predicting the deformation of adjacent tunnels, comprising the following steps: The soil properties, regional geological conditions, and initial constitutive model of the tunnel construction site are obtained. The parameters of the initial constitutive model are set according to the soil properties and regional geological conditions to obtain the initial prediction model. An optimization algorithm is determined, and the initial prediction model is calibrated using the optimization algorithm to obtain multiple parameter combinations. A preset accuracy is obtained, and the multiple parameter combinations are subjected to optimal parameter screening based on the preset accuracy to obtain a target parameter combination. A standard prediction model is obtained based on the target parameter combination. The deformation of the tunnel is predicted using the standard prediction model to obtain the predicted deformation. A warning value estimation neural network is then determined, and the warning deformation is estimated using the warning value estimation neural network to obtain a warning threshold. If the predicted deformation is less than the warning threshold, construction can proceed directly. If the predicted deformation is greater than or equal to the warning threshold, then the engineering performance is obtained, and constraint conditions are constructed based on the engineering performance to obtain the constraint conditions. Then, the support axial force value is calculated based on the warning threshold and the constraint conditions using the standard prediction model to obtain the target support axial force value. Construction instructions are obtained based on the target support axial force value, and these instructions are applied to the construction of the tunnel.

[0007] Optionally, in the method for predicting deformation of adjacent tunnels, the parameters include a first type of parameter and a second type of parameter; The process of acquiring the soil properties, regional geological conditions, and initial constitutive model of the tunnel construction site, and setting the parameters of the initial constitutive model based on the soil properties and regional geological conditions to obtain an initial prediction model, specifically includes: Obtain the soil properties and initial constitutive model of the tunnel construction site, and calculate the first type of parameters of the initial constitutive model based on the soil properties of the construction site to obtain the first parameter configuration; The regional geological conditions of the tunnel are obtained, and the second type of parameters of the initial constitutive model are assigned values ​​according to the regional geological conditions to obtain the second parameter configuration; An initial prediction model is obtained based on the first parameter configuration and the second parameter configuration.

[0008] Optionally, the method for predicting deformation of adjacent tunnels, wherein determining the optimization algorithm involves calibrating the parameters of the initial prediction model using the optimization algorithm to obtain multiple parameter combinations, acquiring a preset accuracy, performing optimal parameter selection on the multiple parameter combinations based on the preset accuracy to obtain a target parameter combination, and obtaining a standard prediction model based on the target parameter combination, specifically including: Determine an optimization algorithm, and use the optimization algorithm to generate candidate parameters for the initial prediction model to obtain multiple parameter combinations; The deformation of the tunnel is evaluated and processed by the initial prediction model based on a combination of multiple parameters to obtain multiple predicted deformation data. Obtain a preset accuracy and the actual deformation data of the tunnel. Based on the preset accuracy, multiple predicted deformation data, and the actual deformation data, perform optimal parameter filtering on multiple parameter combinations to obtain a target parameter combination. Then, obtain a standard prediction model based on the target parameter combination.

[0009] Optionally, the method for predicting the deformation of adjacent tunnels, wherein obtaining a preset accuracy and the actual deformation data of the tunnel, performing optimal parameter filtering on multiple parameter combinations based on the preset accuracy, multiple predicted deformation data, and the actual deformation data to obtain a target parameter combination, and obtaining a standard prediction model based on the target parameter combination, specifically includes: The actual deformation data of the tunnel is obtained, and the residual sum of squares is calculated based on multiple predicted deformation data and the actual deformation data to obtain multiple prediction errors. A preset accuracy is obtained, and multiple parameter combinations are filtered based on the preset accuracy and multiple prediction errors to obtain a target parameter combination. A standard prediction model is then obtained based on the target parameter combination.

[0010] Optionally, the method for predicting the deformation of adjacent tunnels, wherein the deformation of the tunnel is predicted using the standard prediction model to obtain a predicted deformation, and a warning value estimation neural network is determined; a warning deformation is estimated using the warning value estimation neural network to obtain a warning threshold; if the predicted deformation is less than the warning threshold, construction proceeds directly, specifically including: The monitoring points of the tunnel are obtained, and the deformation prediction of the monitoring points is performed using the standard prediction model to obtain the predicted deformation amount. A warning value estimation neural network is determined, real-time engineering data is acquired, and the warning deformation amount is estimated based on the real-time engineering data through the warning value estimation neural network to obtain the warning threshold. If the predicted deformation is less than the warning threshold, construction can proceed directly.

[0011] Optionally, in the method for predicting deformation of adjacent tunnels, the engineering performance includes structural performance and equipment performance; If the predicted deformation is greater than or equal to the warning threshold, then the engineering performance is obtained, and constraint conditions are constructed based on the engineering performance to obtain the constraint conditions. Then, the support axial force value is calculated using the standard prediction model based on the warning threshold and the constraint conditions to obtain the target support axial force value. Specifically, this includes: Obtain the structural performance and the equipment performance, and perform constraint construction processing based on the structural performance and the equipment performance to obtain constraint conditions; The optimization algorithm is used to generate random axial force values, resulting in multiple candidate axial force values. The target support axial force value is obtained by selecting the support axial force value from multiple candidate axial force values ​​based on the constraints and the warning threshold using the standard prediction model.

[0012] Optionally, the method for predicting deformation of adjacent tunnels, wherein obtaining construction instructions based on the target support axial force value and applying the construction instructions to the construction of the tunnel specifically includes: The target support axial force value is processed into a command conversion to obtain a construction command; A supporting force is applied to the tunnel according to the construction instructions, wherein the value of the supporting force is equal to the value of the target supporting axial force.

[0013] Furthermore, to achieve the above objectives, the present invention also provides a deformation prediction system for adjacent tunnels, wherein the deformation prediction system for adjacent tunnels includes: The model configuration module is used to obtain the soil properties, regional geological conditions, and initial constitutive model of the tunnel construction site. Based on the soil properties and regional geological conditions of the construction site, the parameters of the initial constitutive model are set to obtain the initial prediction model. The parameter calibration module is used to determine the optimization algorithm, perform parameter calibration on the initial prediction model through the optimization algorithm to obtain multiple parameter combinations, obtain a preset accuracy, perform optimal parameter screening on the multiple parameter combinations according to the preset accuracy to obtain a target parameter combination, and obtain a standard prediction model according to the target parameter combination. The deformation prediction module is used to predict the deformation of the tunnel using the standard prediction model to obtain the predicted deformation, and to determine the early warning value estimation neural network. The early warning value estimation neural network is used to estimate the early warning deformation to obtain the early warning threshold. If the predicted deformation is less than the early warning threshold, construction can proceed directly. The axial force calculation module is used to obtain engineering performance if the predicted deformation is greater than or equal to the warning threshold, perform constraint condition construction processing based on the engineering performance to obtain constraint conditions, and perform support axial force value calculation processing based on the warning threshold and the constraint conditions through the standard prediction model to obtain the target support axial force value. The instruction application module is used to obtain construction instructions based on the target support axial force value and apply the construction instructions to the construction of the tunnel.

[0014] In addition, to achieve the above objectives, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a deformation prediction program for a neighboring tunnel stored in the memory and executable on the processor, wherein when the deformation prediction program for a neighboring tunnel is executed by the processor, it implements the steps of the deformation prediction method for a neighboring tunnel as described above.

[0015] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a deformation prediction program for a neighboring tunnel, and when the deformation prediction program for a neighboring tunnel is executed by a processor, it implements the steps of the deformation prediction method for a neighboring tunnel as described above.

[0016] In this invention, the soil properties, regional geological conditions, and an initial constitutive model of the tunnel construction site are obtained. The parameters of the initial constitutive model are set according to the soil properties and regional geological conditions to obtain an initial prediction model. An optimization algorithm is determined, and the parameters of the initial prediction model are calibrated using the optimization algorithm to obtain multiple parameter combinations. A preset accuracy is obtained, and the optimal parameter combinations are selected based on the preset accuracy to obtain a target parameter combination. A standard prediction model is then obtained based on the target parameter combination. The deformation of the tunnel is predicted using the standard prediction model to obtain the predicted deformation. A warning value estimation neural network is determined, and the warning deformation is estimated using the warning value estimation neural network to obtain a warning threshold. If the predicted deformation is less than the warning threshold, construction proceeds directly. If the predicted deformation is greater than or equal to the warning threshold, engineering performance is obtained, and constraint conditions are constructed based on the engineering performance to obtain constraint conditions. The support axial force is calculated using the standard prediction model based on the warning threshold and the constraint conditions to obtain a target support axial force value. A construction instruction is obtained based on the target support axial force value and applied to the tunnel construction. This invention uses a standard prediction model to predict the deformation of the tunnel during the next stage of construction, and calculates the target support axial force value based on the predicted deformation. Applying the target support axial force value to the tunnel construction enables timely monitoring of whether the tunnel deformation exceeds the limit and allows for remedial measures. Attached Figure Description

[0017] Figure 1 This is a flowchart of a preferred embodiment of the deformation prediction method for adjacent tunnels of the present invention; Figure 2 This is a system flowchart of the deformation prediction method for adjacent tunnels according to the present invention; Figure 3 This is a physical diagram illustrating the impact of adjacent tunnels on the deformation prediction method of adjacent tunnels according to the present invention. Figure 4 This is an active control optimization flowchart of the deformation prediction method for adjacent tunnels in this invention; Figure 5 This is a structural diagram of a preferred embodiment of the deformation prediction system for adjacent tunnels of the present invention; Figure 6 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] When deep foundation pits are excavated in densely populated urban areas, safety issues often arise with adjacent existing tunnels. The unloading and soil displacement caused by excavation can easily lead to excessive settlement, heave, or elliptic deformation of the tunnel, which may threaten the structural safety and operation of the tunnel in severe cases. Therefore, it is necessary to monitor the deformation of adjacent tunnels during construction.

[0020] Existing methods for predicting the deformation of adjacent tunnels often involve setting up monitoring points during construction and making a one-time prediction using numerical methods such as the finite element method. However, this method is limited by the accuracy of the initial soil and rock parameters and can only passively observe whether the deformation exceeds the limit during construction, making it impossible to monitor whether the tunnel deformation exceeds the limit in a timely manner and take remedial measures.

[0021] To address the aforementioned issues, this invention proposes a method for predicting the deformation of adjacent tunnels. This method uses a standard prediction model to predict the deformation of the tunnel during the next stage of construction, calculates the target support axial force based on the predicted deformation, and applies the target support axial force to the tunnel construction. This allows for timely monitoring of whether the tunnel deformation exceeds the limit and enables remedial measures to be taken.

[0022] The deformation prediction method for adjacent tunnels described in the preferred embodiment of the present invention, such as... Figure 1 As shown, the deformation prediction method for adjacent tunnels includes the following steps: Step S10: Obtain the soil properties, regional geological conditions, and initial constitutive model of the tunnel construction site. Set the parameters of the initial constitutive model according to the soil properties and regional geological conditions to obtain the initial prediction model.

[0023] This invention achieves collaborative modeling through model calibration, prediction and risk assessment, and active control optimization. It constructs a deformation prediction model, combines measured deformation data with deformation control standards to generate construction instructions for optimizing the construction process, and issues construction instructions to achieve timely monitoring and remediation of tunnel deformation.

[0024] like Figure 2 As shown, the steps in this invention to construct a deformation prediction model and optimize the construction process based on the model prediction results include: Step S1: Construct an initial numerical prediction model that includes the tunnel; Step S2: Perform model calibration based on the measured tunnel deformation data from the previous stage; Step S3: Predict and evaluate the tunnel deformation that the next construction step will cause. If the predicted deformation exceeds the limit, proceed to step S4; otherwise, execute the original construction plan. Step S4: Trigger the active control program to perform reverse optimization and solve for the control parameters; Step S5: Perform this construction step, collect new monitoring data, and iterate until the work condition is completed.

[0025] This invention first establishes a three-dimensional numerical analysis model that includes the foundation pit, support structure, adjacent tunnel and surrounding soil. The numerical analysis model can be a finite element model, a finite difference model or a physical information neural network model.

[0026] Specifically, the parameters include a first type of parameter and a second type of parameter; the soil properties of the tunnel construction site and the initial constitutive model are obtained, and the first type of parameter of the initial constitutive model is calculated based on the soil properties of the construction site to obtain the first parameter configuration.

[0027] like Figure 2 As shown, an initial numerical prediction model incorporating the tunnel is first constructed. Specifically, a suitable constitutive model is selected for each soil layer, and an initial set of geotechnical physical and mechanical parameters is assigned based on exploration tests and engineering experience to form the initial prediction model.

[0028] When selecting a suitable constitutive model, the Mohr-Coulomb model is one of the most fundamental models. This model reveals the mechanical behavior of rock materials such as rock, clay, and sand. However, the Mohr-Coulomb model has drawbacks such as fixed stiffness, identical stiffness during loading and unloading, lack of hardening and softening characteristics, and absence of small-strain stiffness features. Therefore, for deep foundation pit excavation involving large deformations and loading / unloading characteristics, a hardening soil model or a hardening soil small-strain model is typically chosen, rather than simply the Mohr-Coulomb model. The hardening soil model and the small-strain hardening model have advantages such as stiffness varying with stress, suitability for large deformations and large stress paths, and the ability to accurately reproduce the true stress-deformation behavior of soil.

[0029] After selecting the constitutive model, basic physical parameters such as soil layer thickness, natural unit weight, and water content are extracted from the geological survey report, and key mechanical parameters such as elastic modulus, cohesion, and internal friction angle are obtained through triaxial tests and consolidation tests.

[0030] Specifically, the spatial distribution information of each soil layer is first extracted from the engineering geological survey report based on borehole columnar sections and geological profiles to determine the thickness of each stratum. Basic physical parameters such as natural unit weight, natural water content, void ratio, and plasticity index are statistically obtained from the geotechnical test results. Compressibility indices such as soil compression modulus and compression coefficient are obtained through indoor consolidation tests, and stress-strain relationships under different stress states are obtained through triaxial shear tests. Key mechanical parameters such as soil elastic modulus, cohesion, internal friction angle, and Poisson's ratio are then determined through inversion. After removing outliers and conducting statistical analysis on multiple sets of test data for the same soil layer, a standardized system of soil physical and mechanical parameters suitable for numerical calculations is formed.

[0031] Further, the regional geological conditions of the tunnel are obtained, and the second type of parameters of the initial constitutive model are assigned values ​​according to the regional geological conditions to obtain the second parameter configuration; the initial prediction model is obtained according to the first parameter configuration and the second parameter configuration.

[0032] For parameters that are difficult to measure precisely (such as the static lateral pressure coefficient), reasonable values ​​and preliminary settings are made by combining regional geological conditions and referring to engineering measured data and standard empirical values ​​under similar geological conditions in the same region, so as to form an operational initial finite element model.

[0033] Step S20: Determine the optimization algorithm, perform parameter calibration on the initial prediction model using the optimization algorithm to obtain multiple parameter combinations, obtain a preset accuracy, perform optimal parameter screening on the multiple parameter combinations according to the preset accuracy to obtain a target parameter combination, and obtain a standard prediction model based on the target parameter combination.

[0034] Specifically, an optimization algorithm is determined, and the initial prediction model is processed by the optimization algorithm to generate candidate parameters, resulting in multiple parameter combinations; the deformation of the tunnel is evaluated by the initial prediction model based on the multiple parameter combinations, resulting in multiple predicted deformation data.

[0035] like Figure 2 As shown, this invention calibrates the initial model based on measured data of tunnel deformation to obtain a standard model that can be used for risk assessment.

[0036] Intelligent optimization algorithms (such as genetic algorithms, particle swarm optimization, or neural network surrogate models) are used to continuously adjust the physical and mechanical parameters of the soil (such as elastic modulus, compression modulus, etc.) within a reasonable range of values, generating multiple candidate parameter combinations.

[0037] The initialization prediction model is debugged and adjusted according to multiple candidate parameter combinations. After each debugging, the deformation of the tunnel is initially evaluated using the debugged prediction model to obtain the corresponding numerical model prediction value.

[0038] Further, the actual deformation data of the tunnel is obtained, and the residual sum of squares is calculated based on the multiple predicted deformation data and the actual deformation data to obtain multiple prediction errors; a preset accuracy is obtained, and multiple parameter combinations are filtered based on the preset accuracy and the multiple prediction errors to obtain a target parameter combination, and a standard prediction model is obtained based on the target parameter combination.

[0039] After one or more foundation pit excavation construction steps are completed, deformation monitoring data of key tunnel sections caused by excavation are collected as on-site measured values.

[0040] The collected deformation monitoring data mainly includes: 1. Tunnel structural displacement: vertical settlement or uplift, and horizontal lateral displacement; 2. Cross-sectional geometric deformation: Tunnel convergence deformation (such as changes in horizontal or vertical diameter) and ellipticity deformation obtained through multi-point monitoring; 3. Ground response: mainly includes the displacement of deep soil around the tunnel (inclinometer data).

[0041] Next, we set an objective function J, which represents the sum of squared residuals between the numerical model predictions and the field measurements; we then substitute each set of numerical model predictions and field measurements recorded earlier into the objective function.

[0042] A preset accuracy is set. When the objective function J decreases to the preset accuracy or stops decreasing, the algorithm stops. The parameter combination at this time is considered to be the "calibrated parameters" that best match the actual soil properties at the current construction site. The model using the "calibrated parameters" is the standard prediction model that is most suitable for deformation prediction.

[0043] Step S30: The deformation of the tunnel is predicted using the standard prediction model to obtain the predicted deformation. A warning value estimation neural network is determined, and the warning deformation is estimated using the warning value estimation neural network to obtain the warning threshold. If the predicted deformation is less than the warning threshold, construction is carried out directly.

[0044] This invention predicts the amount of deformation that will occur during the next stage of construction and conducts a construction risk assessment based on the predicted deformation.

[0045] Specifically, the monitoring points of the tunnel are obtained, and the deformation prediction of the monitoring points is performed using the standard prediction model to obtain the predicted deformation amount.

[0046] like Figure 2 As shown, the tunnel deformation caused by the next construction step is predicted and a risk assessment is conducted using a standard prediction model after parameter calibration.

[0047] like Figure 3 As shown, the next excavation step is simulated on the calibrated model to predict the tunnel deformation (such as settlement, heave, horizontal displacement, and ellipticity deformation) that will result from this step. Specifically, deformation monitoring points set up on the tunnel are acquired. These deformation monitoring points are measurement points set up on buildings or engineering structures to monitor their deformation characteristics in real time. The simulation of the next excavation step is activated in the model (e.g., excavating the next layer of soil or simulating support erection). The maximum deformation increment and cumulative displacement that may occur at each monitoring point of the tunnel under this step are calculated and extracted.

[0048] Furthermore, a warning value estimation neural network is determined, real-time engineering data is acquired, and the warning deformation amount is estimated based on the real-time engineering data through the warning value estimation neural network to obtain the warning threshold.

[0049] After obtaining the deformation prediction results, the predicted deformation is compared with the preset tunnel deformation control standards (such as early warning value and alarm value) to conduct risk level assessment. If the predicted deformation exceeds the limit, active control optimization is initiated to obtain the target support axial force value. The target support axial force value is used to apply corresponding prestress or axial force to the foundation pit support structure to offset the tunnel displacement caused by soil unloading.

[0050] like Figure 4 As shown, the steps of predicting whether the deformation exceeds the limit and triggering the corresponding active control program in this invention include: Step S3.1: Receive the predicted tunnel deformation for the next excavation step, and determine whether the predicted deformation exceeds the warning value based on the deformation control standard database. When the predicted maximum deformation is greater than or equal to the warning value, proceed to the active optimization step S4. When the maximum deformation is less than the warning value, output the original construction parameters. Step S4.1: Set the next support axial force as the optimization variable; Step S4.2: Based on the calibrated geotechnical model, set the optimization model, set the target parameter to find the minimum value of the support axial force, and set the constraint condition to ensure that the tunnel deformation is within the range of the deformation control standard. Step S4.3: Run the intelligent optimization algorithm to solve the optimization problem; Step S4.4: Output the optimal support axial force construction command to guide the actual construction and enter the next monitoring stage.

[0051] The tunnel deformation exceeding the limit early warning value of the present invention can be set by technicians based on experience, or it can be predicted by a neural network model.

[0052] First, a training sample library containing features of multiple working conditions is constructed. Then, a neural network is constructed and trained using geological parameters, deformation rate, construction procedures and environmental risk level as input features and the optimal early warning threshold as the output label. After the model training is completed and verified by the validation set, real-time monitoring data and construction condition information are connected online. The neural network model adaptively infers the early warning threshold for tunnel deformation exceeding the limit under the current working condition.

[0053] Furthermore, if the predicted deformation is less than the warning threshold, construction can proceed directly.

[0054] like Figure 2 and Figure 4 As shown, the predicted deformation is compared with the preset warning value. If the predicted deformation does not exceed the limit, the original construction plan is executed.

[0055] Step S40: If the predicted deformation is greater than or equal to the warning threshold, then the engineering performance is obtained, and constraint conditions are constructed based on the engineering performance to obtain the constraint conditions. The support axial force value is calculated based on the warning threshold and the constraint conditions using the standard prediction model to obtain the target support axial force value.

[0056] When the predicted deformation exceeds the limit, the present invention actively solves for control parameters, which are used in construction to offset the impact of tunnel deformation.

[0057] Specifically, the engineering performance includes structural performance and equipment performance; the structural performance and equipment performance are obtained, and constraint construction processing is performed based on the structural performance and equipment performance to obtain constraint conditions; random axial force value generation processing is performed through the optimization algorithm to obtain multiple candidate axial force values; the target support axial force value is obtained by using the standard prediction model to select support axial force values ​​from the multiple candidate axial force values ​​based on the constraint conditions and the early warning threshold.

[0058] like Figure 2 and Figure 4 As shown, when the predicted deformation exceeds the limit, the active control program is triggered to optimize and solve the control parameters in reverse.

[0059] like Figure 4 As shown, when the predicted deformation exceeds the limit, the axial force of the next support is set as the optimization variable, and the minimum axial force value that can just suppress the tunnel deformation below the warning line is solved; an optimization model containing the objective function and constraints is constructed; and an intelligent optimization algorithm is run according to the optimization model to solve the optimization problem and obtain the optimal support axial force value.

[0060] First, constraints are constructed based on structural performance and equipment capabilities. The factors considered in the constraints include: 1. Bearing capacity of the supporting structure: The axial force must not exceed the design value of the material strength of the steel or concrete support to prevent the support itself from becoming unstable and failing; 2. Limits of construction equipment: The axial force must not exceed the maximum rated jacking force of the hydraulic jacks on site; 3. Safety of the foundation pit itself: Excessive support force may cause the retaining wall to deform improperly to the outside of the foundation pit.

[0061] Next, an optimization model is constructed. Specifically, the value of the support axial force is set as an optimization function, ensuring that the value of the optimization function simultaneously satisfies all constraints and that the predicted maximum tunnel deformation is exactly equal to or slightly less than the warning value. "Slightly less than" refers to a safety redundancy range, the specific range of which is set by engineers based on the tunnel's characteristics, generally between 85% and 95% of the warning value. The goal is to prevent deformation from reaching the warning line without excessively wasting support resources. When setting the safety redundancy range, different ranges can be used for standard and risk sections of the tunnel, and different redundancies can also be set for different measuring points on the same cross-section, achieving refined control.

[0062] When setting up a safety redundancy range, in addition to having engineers set it based on experience, a neural network can be trained by combining historical deformation exceeding cases of this project or region, and the safety factor can be automatically corrected through the neural network model.

[0063] The intelligent optimization algorithm is invoked again, and the calibrated prediction model is run multiple times in the background. Each time the intelligent optimization algorithm is invoked, it randomly generates a set of axial forces. This set of axial forces is imported into the standard prediction model, and the algorithm will discard axial force values ​​that still do not meet the constraints after applying the axial force. After multiple iterations, the minimum axial force value that can suppress the tunnel deformation below the warning line is finally determined as the optimal solution.

[0064] Step S50: Obtain construction instructions based on the target support axial force value, and apply the construction instructions to the construction of the tunnel.

[0065] This invention transforms the control parameters output by the active control optimization module into construction parameter commands for application during construction, and iteratively executes the steps of model calibration, prediction, evaluation, and control parameter optimization until construction is completed.

[0066] like Figure 2 As shown, the support axial force value calculated based on the predicted deformation will serve as a clear and quantifiable construction instruction for the next construction step.

[0067] Specifically, the target support axial force value is processed into a construction command; a support force is applied to the tunnel according to the construction command, wherein the value of the support force is equal to the target support axial force value.

[0068] The system sends the specific axial force value to the control terminal of the on-site construction team; the construction personnel operate the hydraulic pump station and apply the corresponding prestress to the steel support through the jacks; when the jack reading reaches the optimized value required by the system, physical locking is performed, thereby offsetting the tunnel displacement caused by soil unloading through active support force.

[0069] When applying axial force, this invention does not apply it to the target value all at once, but in multiple stages. After each stage is completed, the tunnel deformation feedback is read before executing the next stage, avoiding impact loading and overshoot. During the application process, the tunnel deformation and axial force are monitored in real time. If the deviation exceeds the limit, the command is immediately recalculated and updated. Since the equipment at different construction sites is different, this invention corrects the target axial force according to the pressure-flow characteristics of different jacks and oil pumps, generating accurate pressure commands adapted to the equipment on site, thereby improving the accuracy of axial force control.

[0070] like Figure 2 As shown, after the current construction is completed, the system uses the newly collected tunnel deformation data to perform a new round of model calibration, prediction, evaluation and control parameter optimization. This iterative process continues until the entire foundation pit excavation is completed.

[0071] After predicting the deformation of the tunnel, this invention can be used not only to calculate the axial force applied to the foundation pit support structure, but also to calculate support parameters such as vertical spacing of supports, longitudinal spacing of supports, number of temporary supports, and timing of prestressing procedures; or to calculate grouting parameters such as grouting pressure, grouting diffusion radius, and grouting hole spacing; or to calculate other deformation control measures such as anchor pipe, anchor length and inclination angle, tension control force, and backfill compaction.

[0072] The technical effects of this invention include: 1. Traditionally, numerical methods such as the finite element method are used for one-time prediction before construction, and monitoring points are set up during construction. The accuracy of this method is limited by the accuracy of the initial soil and rock parameters, and during construction, deformation can only be passively observed. When monitoring data alarms, the situation is often already reactive, with costly and ineffective remedial measures. This invention uses a predictive model to accurately predict the tunnel deformation after the next stage of construction and calculates the corresponding axial force value based on the deformation to generate construction instructions. This allows for timely monitoring of tunnel deformation exceeding limits and appropriate remedial measures.

[0073] 2. If tunnel deformation approaches the warning value during construction, in order to effectively control further deformation of the surrounding rock and support structure, and ensure construction safety and the stability of the surrounding environment, engineering projects typically adopt active and passive control measures such as increasing the prestress of steel or concrete supports, and implementing grouting reinforcement of the surrounding rock or strata. The parameters of these measures (such as the magnitude of the support force and the amount of grout) largely rely on engineering experience and lack quantitative basis, which may lead to insufficient reinforcement or excessive waste. This invention, based on high-precision predicted tunnel deformation, takes the strict compliance of tunnel deformation with safety control standards as its optimization objective. It inversely calculates the optimal support axial force value that precisely offsets the tunnel deformation effect. Under the premise of effectively suppressing deformation and meeting structural and environmental safety requirements, it minimizes the cost of support application force, engineering materials, and energy consumption, avoiding resource waste caused by excessive reinforcement.

[0074] 3. Existing back analysis methods are mainly used for post-hoc interpretation of deformations and calibration of soil parameters. However, these methods often stop at the parameter update stage and fail to quickly and directly use the high-precision parameters obtained from the back analysis for control parameter optimization. They generally lack real-time decision-making capabilities for key active control parameters, making it difficult to form a complete link from parameter identification to control commands and guide the next stage of construction. This invention discloses a closed-loop intelligent method that integrates deformation prediction, risk assessment, parameter back analysis, and active control. Through data exchange and iterative optimization among multiple modules, it achieves dynamic and precise control of "prediction-assessment-intervention" of deformation of adjacent tunnels.

[0075] This invention uses a standard prediction model to predict the deformation of the tunnel during the next stage of construction, and calculates the target support axial force value based on the predicted deformation. Applying the target support axial force value to the tunnel construction enables timely monitoring of whether the tunnel deformation exceeds the limit and allows for remedial measures.

[0076] Furthermore, such as Figure 5 As shown, based on the above-mentioned deformation prediction method for adjacent tunnels, the present invention also provides a deformation prediction system for adjacent tunnels, wherein the deformation prediction system for adjacent tunnels includes: The model configuration module 51 is used to obtain the soil properties, regional geological conditions and initial constitutive model of the tunnel construction site, and to set the parameters of the initial constitutive model according to the soil properties and regional geological conditions of the construction site to obtain the initial prediction model. The parameter calibration module 52 is used to determine the optimization algorithm, perform parameter calibration on the initial prediction model through the optimization algorithm to obtain multiple parameter combinations, obtain a preset accuracy, perform optimal parameter screening on the multiple parameter combinations according to the preset accuracy to obtain a target parameter combination, and obtain a standard prediction model according to the target parameter combination. The deformation prediction module 53 is used to predict the deformation of the tunnel using the standard prediction model to obtain the predicted deformation, and to determine the early warning value estimation neural network. The early warning value estimation neural network is used to estimate the early warning deformation to obtain the early warning threshold. If the predicted deformation is less than the early warning threshold, construction is carried out directly. The axial force calculation module 54 is used to obtain engineering performance if the predicted deformation is greater than or equal to the warning threshold, perform constraint condition construction processing based on the engineering performance to obtain constraint conditions, and perform support axial force value calculation processing based on the warning threshold and the constraint conditions through the standard prediction model to obtain the target support axial force value. The instruction application module 55 is used to obtain construction instructions based on the target support axial force value and apply the construction instructions to the construction of the tunnel.

[0077] Furthermore, such as Figure 6 As shown, based on the above-mentioned deformation prediction method and system for adjacent tunnels, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 6 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0078] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc. Further, the memory 20 may include both internal and external storage devices. The memory 20 is used to store application software and various types of data installed on the terminal, such as program code installed on the terminal. The memory 20 may also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a deformation prediction program 40 for adjacent tunnels, which can be executed by the processor 10 to implement the deformation prediction method for adjacent tunnels in this application.

[0079] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing the deformation prediction method of the adjacent tunnel.

[0080] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface.

[0081] In one embodiment, when the processor 10 executes the deformation prediction program 40 for adjacent tunnels in the memory 20, the following steps are performed: The soil properties, regional geological conditions, and initial constitutive model of the tunnel construction site are obtained. The parameters of the initial constitutive model are set according to the soil properties and regional geological conditions to obtain the initial prediction model. An optimization algorithm is determined, and the initial prediction model is calibrated using the optimization algorithm to obtain multiple parameter combinations. A preset accuracy is obtained, and the multiple parameter combinations are subjected to optimal parameter screening based on the preset accuracy to obtain a target parameter combination. A standard prediction model is obtained based on the target parameter combination. The deformation of the tunnel is predicted using the standard prediction model to obtain the predicted deformation. A warning value estimation neural network is then determined, and the warning deformation is estimated using the warning value estimation neural network to obtain a warning threshold. If the predicted deformation is less than the warning threshold, construction can proceed directly. If the predicted deformation is greater than or equal to the warning threshold, then the engineering performance is obtained, and constraint conditions are constructed based on the engineering performance to obtain the constraint conditions. Then, the support axial force value is calculated based on the warning threshold and the constraint conditions using the standard prediction model to obtain the target support axial force value. Construction instructions are obtained based on the target support axial force value, and these instructions are applied to the construction of the tunnel.

[0082] The parameters include a first type of parameter and a second type of parameter; The process of acquiring the soil properties, regional geological conditions, and initial constitutive model of the tunnel construction site, and setting the parameters of the initial constitutive model based on the soil properties and regional geological conditions to obtain an initial prediction model, specifically includes: Obtain the soil properties and initial constitutive model of the tunnel construction site, and calculate the first type of parameters of the initial constitutive model based on the soil properties of the construction site to obtain the first parameter configuration; The regional geological conditions of the tunnel are obtained, and the second type of parameters of the initial constitutive model are assigned values ​​according to the regional geological conditions to obtain the second parameter configuration; An initial prediction model is obtained based on the first parameter configuration and the second parameter configuration.

[0083] The determination of the optimization algorithm involves using the optimization algorithm to perform parameter calibration on the initial prediction model, obtaining multiple parameter combinations, acquiring a preset accuracy, performing optimal parameter selection on the multiple parameter combinations based on the preset accuracy, obtaining a target parameter combination, and obtaining a standard prediction model based on the target parameter combination. Specifically, this includes: Determine an optimization algorithm, and use the optimization algorithm to generate candidate parameters for the initial prediction model to obtain multiple parameter combinations; The deformation of the tunnel is evaluated and processed by the initial prediction model based on a combination of multiple parameters to obtain multiple predicted deformation data. Obtain a preset accuracy and the actual deformation data of the tunnel. Based on the preset accuracy, multiple predicted deformation data, and the actual deformation data, perform optimal parameter filtering on multiple parameter combinations to obtain a target parameter combination. Then, obtain a standard prediction model based on the target parameter combination.

[0084] The process of obtaining a preset precision and the actual deformation data of the tunnel, performing optimal parameter filtering on multiple parameter combinations based on the preset precision, multiple predicted deformation data, and the actual deformation data to obtain a target parameter combination, and obtaining a standard prediction model based on the target parameter combination, specifically includes: The actual deformation data of the tunnel is obtained, and the residual sum of squares is calculated based on multiple predicted deformation data and the actual deformation data to obtain multiple prediction errors. A preset accuracy is obtained, and multiple parameter combinations are filtered based on the preset accuracy and multiple prediction errors to obtain a target parameter combination. A standard prediction model is then obtained based on the target parameter combination.

[0085] Specifically, the process of predicting the deformation of the tunnel using the standard prediction model to obtain the predicted deformation, determining a warning value estimation neural network, estimating the warning deformation using the warning value estimation neural network to obtain a warning threshold, and proceeding with construction directly if the predicted deformation is less than the warning threshold, includes: The monitoring points of the tunnel are obtained, and the deformation prediction of the monitoring points is performed using the standard prediction model to obtain the predicted deformation amount. A warning value estimation neural network is determined, real-time engineering data is acquired, and the warning deformation amount is estimated based on the real-time engineering data through the warning value estimation neural network to obtain the warning threshold. If the predicted deformation is less than the warning threshold, construction can proceed directly.

[0086] The engineering performance includes structural performance and equipment performance; If the predicted deformation is greater than or equal to the warning threshold, then the engineering performance is obtained, and constraint conditions are constructed based on the engineering performance to obtain the constraint conditions. Then, the support axial force value is calculated using the standard prediction model based on the warning threshold and the constraint conditions to obtain the target support axial force value. Specifically, this includes: Obtain the structural performance and the equipment performance, and perform constraint construction processing based on the structural performance and the equipment performance to obtain constraint conditions; The optimization algorithm is used to generate random axial force values, resulting in multiple candidate axial force values. The target support axial force value is obtained by selecting the support axial force value from multiple candidate axial force values ​​based on the constraints and the warning threshold using the standard prediction model.

[0087] Specifically, the step of obtaining construction instructions based on the target support axial force value and applying the construction instructions to the tunnel construction includes: The target support axial force value is processed into a command conversion to obtain a construction command; A supporting force is applied to the tunnel according to the construction instructions, wherein the value of the supporting force is equal to the value of the target supporting axial force.

[0088] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a deformation prediction program for a neighboring tunnel, and the deformation prediction program for a neighboring tunnel, when executed by a processor, implements the steps of the deformation prediction method for a neighboring tunnel as described above.

[0089] In summary, this invention provides a method, system, and terminal for predicting deformation of adjacent tunnels. The method includes: acquiring the soil properties, regional geological conditions, and an initial constitutive model of the tunnel construction site; setting the parameters of the initial constitutive model based on the soil properties and regional geological conditions to obtain an initial prediction model; determining an optimization algorithm; performing parameter calibration on the initial prediction model using the optimization algorithm to obtain multiple parameter combinations; obtaining a preset accuracy; performing optimal parameter selection on the multiple parameter combinations based on the preset accuracy to obtain a target parameter combination; and obtaining a standard prediction model based on the target parameter combination; and using the standard prediction model to predict the deformation of the adjacent tunnel. The deformation of the tunnel is predicted to obtain a predicted deformation amount, and a warning value estimation neural network is determined. This network is then used to estimate the predicted deformation amount, resulting in a warning threshold. If the predicted deformation amount is less than the warning threshold, construction proceeds directly. If the predicted deformation amount is greater than or equal to the warning threshold, engineering performance is obtained, and constraints are constructed based on these performance conditions. A standard prediction model is then used to calculate the target support axial force value based on the warning threshold and the constraints. Based on this target support axial force value, a construction instruction is generated and applied to the tunnel construction. This invention uses a standard prediction model to predict the tunnel deformation during the next stage of construction, calculates the target support axial force value based on the predicted deformation, and applies the target support axial force value to the tunnel construction. This allows for timely monitoring of whether the tunnel deformation exceeds limits and enables remedial measures.

[0090] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0091] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.

[0092] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method of deformation prediction in proximity to a tunnel, characterized by, The deformation prediction method for the adjacent tunnel includes: The soil properties, regional geological conditions, and initial constitutive model of the tunnel construction site are obtained. The parameters of the initial constitutive model are set according to the soil properties and regional geological conditions to obtain the initial prediction model. An optimization algorithm is determined, and the initial prediction model is calibrated using the optimization algorithm to obtain multiple parameter combinations. A preset accuracy is obtained, and the multiple parameter combinations are subjected to optimal parameter screening based on the preset accuracy to obtain a target parameter combination. A standard prediction model is then obtained based on the target parameter combination. The deformation of the tunnel is predicted using the standard prediction model to obtain the predicted deformation. A warning value estimation neural network is then determined, and the warning deformation is estimated using the warning value estimation neural network to obtain a warning threshold. If the predicted deformation is less than the warning threshold, construction can proceed directly. If the predicted deformation is greater than or equal to the warning threshold, then the engineering performance is obtained, and constraint conditions are constructed based on the engineering performance to obtain the constraint conditions. Then, the support axial force value is calculated based on the warning threshold and the constraint conditions using the standard prediction model to obtain the target support axial force value. Construction instructions are obtained based on the target support axial force value, and these instructions are applied to the construction of the tunnel. The parameters include first-type parameters and second-type parameters; The process of acquiring the soil properties, regional geological conditions, and initial constitutive model of the tunnel construction site, and setting the parameters of the initial constitutive model based on the soil properties and regional geological conditions to obtain an initial prediction model, specifically includes: Obtain the soil properties and initial constitutive model of the tunnel construction site, and calculate the first type of parameters of the initial constitutive model based on the soil properties of the construction site to obtain the first parameter configuration; The regional geological conditions of the tunnel are obtained, and the second type of parameters of the initial constitutive model are assigned values ​​according to the regional geological conditions to obtain the second parameter configuration; An initial prediction model is obtained based on the first parameter configuration and the second parameter configuration; The determination of the optimization algorithm involves calibrating the parameters of the initial prediction model using the optimization algorithm to obtain multiple parameter combinations, acquiring a preset accuracy, performing optimal parameter selection on the multiple parameter combinations based on the preset accuracy to obtain a target parameter combination, and obtaining a standard prediction model based on the target parameter combination. Specifically, this includes: Determine an optimization algorithm, and use the optimization algorithm to generate candidate parameters for the initial prediction model to obtain multiple parameter combinations; The deformation of the tunnel is evaluated and processed by the initial prediction model based on a combination of multiple parameters to obtain multiple predicted deformation data. Obtain a preset accuracy and the actual deformation data of the tunnel; perform optimal parameter filtering on multiple parameter combinations based on the preset accuracy, multiple predicted deformation data and the actual deformation data to obtain a target parameter combination; and obtain a standard prediction model based on the target parameter combination. The engineering performance includes structural performance and equipment performance; If the predicted deformation is greater than or equal to the warning threshold, then the engineering performance is obtained, and constraint conditions are constructed based on the engineering performance to obtain the constraint conditions. Then, the support axial force value is calculated using the standard prediction model based on the warning threshold and the constraint conditions to obtain the target support axial force value. Specifically, this includes: Obtain the structural performance and the equipment performance, and perform constraint construction processing based on the structural performance and the equipment performance to obtain constraint conditions; The optimization algorithm is used to generate random axial force values, resulting in multiple candidate axial force values. The target support axial force value is obtained by selecting the support axial force value from multiple candidate axial force values ​​based on the constraints and the warning threshold using the standard prediction model.

2. The method of claim 1, wherein, The process of obtaining a preset precision and the actual deformation data of the tunnel, performing optimal parameter filtering on multiple parameter combinations based on the preset precision, multiple predicted deformation data, and the actual deformation data to obtain a target parameter combination, and obtaining a standard prediction model based on the target parameter combination, specifically includes: The actual deformation data of the tunnel is obtained, and the residual sum of squares is calculated based on multiple predicted deformation data and the actual deformation data to obtain multiple prediction errors. A preset accuracy is obtained, and multiple parameter combinations are filtered based on the preset accuracy and multiple prediction errors to obtain a target parameter combination. A standard prediction model is then obtained based on the target parameter combination.

3. The deformation prediction method for adjacent tunnels according to claim 1, characterized in that, The process involves predicting the tunnel deformation using the standard prediction model to obtain the predicted deformation, determining a warning value estimation neural network, estimating the warning deformation using the warning value estimation neural network to obtain a warning threshold, and proceeding with construction directly if the predicted deformation is less than the warning threshold. Specifically, this includes: The monitoring points of the tunnel are obtained, and the deformation prediction of the monitoring points is performed using the standard prediction model to obtain the predicted deformation amount. A warning value estimation neural network is determined, real-time engineering data is acquired, and the warning deformation amount is estimated based on the real-time engineering data through the warning value estimation neural network to obtain the warning threshold. If the predicted deformation is less than the warning threshold, construction can proceed directly.

4. The deformation prediction method for adjacent tunnels according to claim 1, characterized in that, The process of obtaining construction instructions based on the target support axial force value and applying these instructions to the tunnel construction specifically includes: The target support axial force value is processed into a command conversion to obtain a construction command; A supporting force is applied to the tunnel according to the construction instructions, wherein the value of the supporting force is equal to the value of the target supporting axial force.

5. A deformation prediction system for adjacent tunnels, wherein the deformation prediction system for adjacent tunnels is used to implement the deformation prediction method for adjacent tunnels according to any one of claims 1-4, characterized in that, The deformation prediction system for the adjacent tunnel includes: The model configuration module is used to obtain the soil properties, regional geological conditions, and initial constitutive model of the tunnel construction site. Based on the soil properties and regional geological conditions of the construction site, the parameters of the initial constitutive model are set to obtain the initial prediction model. The parameter calibration module is used to determine the optimization algorithm, perform parameter calibration on the initial prediction model through the optimization algorithm to obtain multiple parameter combinations, obtain a preset accuracy, perform optimal parameter screening on the multiple parameter combinations according to the preset accuracy to obtain a target parameter combination, and obtain a standard prediction model according to the target parameter combination. The deformation prediction module is used to predict the deformation of the tunnel using the standard prediction model to obtain the predicted deformation, and to determine the early warning value estimation neural network. The early warning value estimation neural network is used to estimate the early warning deformation to obtain the early warning threshold. If the predicted deformation is less than the early warning threshold, construction can proceed directly. The axial force calculation module is used to obtain engineering performance if the predicted deformation is greater than or equal to the warning threshold, perform constraint condition construction processing based on the engineering performance to obtain constraint conditions, and perform support axial force value calculation processing based on the warning threshold and the constraint conditions through the standard prediction model to obtain the target support axial force value. The instruction application module is used to obtain construction instructions based on the target support axial force value and apply the construction instructions to the construction of the tunnel.

6. A terminal, characterized in that, The terminal includes a memory, a processor, and a deformation prediction program for a neighboring tunnel stored in the memory and executable on the processor. When the processor executes the deformation prediction program for a neighboring tunnel, it implements the steps of the deformation prediction method for a neighboring tunnel as described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a deformation prediction program for a neighboring tunnel, which, when executed by a processor, implements the steps of the deformation prediction method for a neighboring tunnel as described in any one of claims 1-4.

Images