Vertical micro-deformation prediction method and system for subway structure in subway double-sided extension

Abstract

In order to improve the prediction accuracy of the vertical micro-deformation of the subway structure during the two-sided extension of the subway, the present invention proposes a method and system for predicting the vertical micro-deformation of the subway structure during the two-sided extension of the subway, which relates to the field of deformation prediction technology. In view of the problem of low prediction accuracy of traditional numerical simulation methods, the present method selects key locations around the subway station to conduct physical test tests, obtains physical test test parameters and inputs them into the numerical simulation model to obtain the simulated vertical deformation of the soil. Actual excavation is carried out in the adjacent strata to obtain the actual vertical deformation of the soil, and a correction model is constructed to correct the simulated vertical deformation of the soil to obtain the simulated optimal vertical deformation of the soil. By comparing the simulated optimal vertical deformation of the soil with the actual vertical deformation, the correction coefficient is calculated to correct the vertical micro-deformation of the subway station to obtain the final predicted value. The present method can effectively improve the prediction accuracy of the vertical micro-deformation of the subway station and provide reliable technical support for the safe construction of subway projects.

Description

Technical Field

[0001] The present application relates to the field of deformation prediction technology, and in particular to a method and system for predicting vertical micro-deformation of subway structures during double-sided subway expansion. Background Art

[0002] With the acceleration of urbanization, the development and utilization of underground space is increasing. As an important part of urban public transportation, the construction and operation safety of subways have become the focus of people's attention. The expansion of both sides of subway stations often involves large-scale underground excavation projects, which will affect the soil and further cause vertical micro-deformation of subway structures (including subway stations and shield sections). The vertical micro-deformation of the soil will affect the operational safety of the operating subway. Therefore, accurately predicting the vertical micro-deformation of subway stations is crucial to ensure construction safety and project quality.

[0003] At present, numerical simulation methods are widely used to predict the vertical micro-deformation of subway structures during the expansion of both sides of subway stations. This method can better reflect the interaction between soil and subway stations by establishing a three-dimensional numerical model, thereby predicting the deformation of soil during construction. However, although the numerical simulation method has certain advantages, it also has some technical defects: soft soil has the characteristics of high water content, underconsolidation, low bearing capacity, high compressibility and large settlement. In general, soft soil is usually treated without being used as a natural foundation. When soft soil appears in the form of a building environment, there are certain engineering risks. Especially when it is disturbed by construction, its mechanical behavior will change. In addition, the engineering properties of soft soil in different regions vary greatly due to different causes and have very obvious regional characteristics, so it is difficult to obtain a universal constitutive model. In addition, the parameters of most constitutive models are unclear and need to be determined through unconventional tests, resulting in differences between numerical simulation results and actual deformation. The heterogeneity and discreteness of the actual strata have a great influence on soil deformation, but it is difficult to accurately reflect these complex geological characteristics in numerical simulation, which further increases the difficulty of prediction. Summary of the invention

[0004] The purpose of this application is to provide a method and system for predicting vertical micro-deformation of subway structures in subway double-sided extension, so as to realize the vertical micro-deformation prediction of large subway stations.

[0005] On the one hand, according to the embodiment of the present application, a method for predicting vertical micro-deformation of a subway structure with two-sided subway expansion is proposed. Step 1: select key locations around the subway station for sampling, and perform physical test tests on the stress paths during excavation of the key locations to obtain physical test test results: physical test test parameters;

[0006] Step 2: Based on the test results of the physical experiment, a numerical simulation model is selected, and the model parameters of the numerical simulation model are initialized. Then, the test parameters of the physical experiment are input into the numerical simulation model to obtain the simulation results of the numerical simulation model: the vertical deformation of the soil and the subway station, and the simulated vertical deformation of the soil is obtained according to the vertical deformation of the soil and the subway station;

[0007] Step 3: According to the simulation results of the numerical simulation model, excavation is performed in the adjacent strata around the subway station to obtain the actual vertical deformation of the soil, and the vertical micro deformation of the subway station is obtained according to the actual vertical deformation of the soil;

[0008] Step 4, constructing a correction model, correcting the simulated vertical deformation of the soil, and obtaining the optimal simulated vertical deformation of the soil;

[0009] Step 5, obtaining a correction coefficient of the vertical micro-deformation of the subway station according to the simulated optimal vertical deformation of the soil and the actual vertical deformation of the soil;

[0010] Step 6: Obtain the predicted value of the vertical micro-deformation of the subway station according to the correction coefficient and the vertical micro-deformation of the subway station.

[0011] According to one aspect of the embodiment of the present application, step 4, constructing a correction model, correcting the simulated vertical deformation of the soil, and obtaining the simulated optimal vertical deformation of the soil, specifically includes the following steps:

[0012] Step 41, obtaining a training set of the revised model;

[0013] Step 42, constructing an iterative function according to the training set;

[0014] Step 43, constructing a threshold condition for the iterative function;

[0015] Step 44, input the training set into the correction model for training, and obtain a trained correction model;

[0016] Step 45, inputting the actual vertical deformation of the soil and the simulated vertical deformation of the soil into the trained correction model, and outputting the first optimized vertical deformation of the simulated soil;

[0017] Step S46, calculating the first prediction error between the first optimized vertical deformation of the soil simulation and the training set;

[0018] Step S47, compare the first prediction error with the threshold condition. If the first prediction error is less than or equal to the threshold condition, the output first optimized vertical deformation of soil simulation is the optimal vertical deformation of soil simulation; if the first prediction error is greater than the threshold condition, iterative processing is performed to obtain the optimal vertical deformation of soil simulation to realize the correction of the vertical deformation of soil simulation.

[0019] According to one aspect of an embodiment of the present application, the training set includes several sample sets, each of which includes: the actual historical vertical deformation of the soil, the simulated historical vertical deformation of the soil, the simulated historical optimal vertical deformation of the soil, the historical model parameters of the numerical simulation model, the historical stress path of the excavation process, the historical coordinate information of the sampling location, the historical parameters of physical test tests, the historical parameters of soil characteristics, and the historical parameters of adjacent stratum structure.

[0020] According to one aspect of an embodiment of the present application, if the first prediction error is greater than a threshold condition, an iterative process is performed to obtain an optimal vertical deformation of soil simulation, specifically including the following steps:

[0021] Step 471, re-perform physical test on the stress path during the excavation process of the key position to obtain physical test optimization parameters, optimize the model parameters after the initialization processing, input the physical test optimization parameters into the numerical simulation model, and obtain the optimized vertical deformation of the soil and the subway station according to the numerical simulation model, and obtain the optimized simulated vertical deformation of the soil according to the optimized vertical deformation;

[0022] Step 472, adding the optimized simulated vertical deformation of the soil, the optimized model parameters, the physical test optimization parameters, and the optimized simulated vertical deformation of the soil as sub-sample sets to the training set to obtain an optimized training set;

[0023] Step S473, retraining the correction model using the optimized training set to obtain a trained optimized correction model;

[0024] Step S474, inputting the actual vertical deformation of the soil and the optimized simulated vertical deformation of the soil into the trained optimization correction model, and outputting the simulated second optimized vertical deformation of the soil;

[0025] Step S475, calculate the second prediction error between the second optimized vertical deformation of soil simulation and the optimized training set. If the second prediction error is less than or equal to the threshold condition, the second optimized vertical deformation of soil simulation is the optimal vertical deformation of soil simulation; if the second prediction error is greater than the threshold condition, return to execute steps 472-475 for iterative processing. When the second prediction error after iterative processing is less than or equal to the threshold condition or the number of iterations reaches the maximum number of iterations, stop the iteration. The second optimized vertical deformation of soil simulation after the last iterative processing is the optimal vertical deformation of soil simulation.

[0026] According to one aspect of the embodiment of the present application, the iterative function is:

[0027] ;

[0028] ;

[0029] Where L represents the iterative function, represents the optimal vertical deformation of the soil simulation history of the i-th sample set, represents the optimal vertical deformation of soil simulation predicted by the modified model, n represents the maximum number of sample sets, M represents the modified model, D represents the optimized training set, represents the iteration condition, represents the threshold condition, F represents stopping the iteration operation, and T represents the iteration operation.

[0030] According to one aspect of an embodiment of the present application, the training set is obtained from historical soil excavation cases, the historical optimal vertical deformation of the soil simulation is used as the output label of the correction model, the input of the correction model is the actual vertical deformation of the soil and the simulated vertical deformation of the soil, and the output of the correction model is the optimal vertical deformation of the soil simulation.

[0031] According to one aspect of an embodiment of the present application, the correction model is a deep learning model.

[0032] According to one aspect of the embodiment of the present application, the calculation formula of the correction coefficient is:

[0033] ;

[0034] Where N is the correction factor, Indicates the actual vertical deformation of the soil. It represents the optimal vertical deformation of soil simulation output by the last iteration of the numerical simulation model.

[0035] According to one aspect of the embodiment of the present application, the calculation formula for the predicted value of the vertical micro-deformation of the subway station is:

[0036] ;

[0037] In the formula, Y represents the predicted value of vertical micro-deformation of the subway station, N represents the correction coefficient, Indicates the vertical micro deformation of the subway station.

[0038] On the other hand, the embodiment of the present application proposes a vertical micro-deformation prediction system for a large subway station, comprising the following modules:

[0039] Physical test module: select key locations around the subway station for sampling, and conduct physical test tests on the stress paths during excavation at key locations to obtain physical test results: physical test parameters;

[0040] Soil simulation vertical deformation acquisition module: connected with the physical test module, based on the physical test results, selects the numerical simulation model, initializes the model parameters of the numerical simulation model, and then inputs the physical test parameters into the numerical simulation model to obtain the simulation results of the numerical simulation model: the vertical deformation of the soil and the subway station, and obtains the simulated vertical deformation of the soil according to the vertical deformation of the soil and the subway station;

[0041] The actual vertical deformation acquisition module of the soil is connected to the simulated vertical deformation acquisition module of the soil, and is used to excavate the adjacent strata around the subway station according to the simulation results of the numerical simulation model to obtain the actual vertical deformation of the soil, and obtain the vertical micro deformation of the subway station according to the actual vertical deformation of the soil;

[0042] Correction model: connected with the soil simulation vertical deformation acquisition module, used to construct a correction model, correct the soil simulation vertical deformation, and obtain the optimal soil simulation vertical deformation;

[0043] Correction coefficient calculation module: connected with the correction model, used to obtain the correction coefficient according to the optimal vertical deformation of the soil simulation and the actual vertical deformation of the soil;

[0044] Prediction value calculation module: connected to the correction coefficient calculation module, used to obtain the prediction value of the vertical micro-deformation of the subway station according to the correction coefficient and the vertical micro-deformation of the subway station.

[0045] In summary, the beneficial technical effects of this application are:

[0046] The present invention first obtains the simulated vertical deformation of the soil through a numerical simulation model, then performs on-site excavation to obtain the actual vertical deformation of the soil and the vertical micro-deformation of the subway station, and obtains the optimal vertical deformation of the soil simulation by using the prediction results of the correction model based on the actual vertical deformation of the excavated soil and the simulated vertical deformation of the soil by numerical simulation. Finally, a correction coefficient is obtained based on the optimal vertical deformation of the soil simulation and the actual vertical deformation of the soil, and the vertical micro-deformation of the subway station is corrected by the correction coefficient to obtain the predicted value of the vertical micro-deformation of the subway station. By combining the data obtained by numerical simulation and on-site excavation, the actual working conditions can be more accurately reflected, and the simulated vertical deformation of the soil can be effectively corrected, thereby improving the prediction accuracy of the vertical micro-deformation of the subway station and reducing the difficulty of prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a flow chart of the method for predicting vertical micro-deformation of subway structure for subway double-sided extension in this application;

[0048] Figure 2 It is a structural diagram of a vertical micro-deformation prediction system for a large subway station according to another embodiment of the present application. DETAILED DESCRIPTION

[0049] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.

[0050] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0051] In the embodiments of the present application, the same reference numerals represent the same components, and for the sake of brevity, detailed descriptions of the same components are omitted in different embodiments. It should be understood that the thickness, length, width and other dimensions of various components in the embodiments of the present application shown in the drawings, as well as the overall thickness, length, width and other dimensions of the integrated device are only exemplary descriptions and should not constitute any limitation to the present application.

[0052] The term “plurality” used in this application refers to two or more (including two).

[0053] On the one hand, the following combination Figure 1 to Figure 2 A method and system for predicting vertical micro-deformation of a subway structure during double-sided extension of a subway according to an embodiment of the present application are described in detail.

[0054] See attached Figure 1 -Attached Figure 2 The vertical micro-deformation prediction method of subway structure for subway double-side extension includes the following steps:

[0055] Step 1: select key locations around the subway station for sampling, and conduct physical test tests on the stress paths during excavation at the key locations to obtain physical test results: physical test parameters.

[0056] According to geological conditions, underground structure design and construction experience, the locations around the subway station that have the greatest impact on the stability of the subway station are selected as key locations, and the soil is excavated at this location to install the complex and enhance the stability of the subway station. Key locations include but are not limited to: station entrances and exits, transfer passages, tunnel intersections, etc. Excavate the selected key locations, obtain soil samples, and design a series of physical tests, such as triaxial compression tests, direct shear tests, static load tests, etc., to simulate the stress state under actual working conditions. During the test, record the stress, strain, strength and other related parameters of the soil. Physical test test parameters include but are not limited to: compressive strength, tensile strength, elastic modulus, Poisson's ratio, permeability, saturation, water content, plasticity index, liquid limit, solid limit, etc. of the soil.

[0057] The stress path refers to a series of loading states that the soil material experiences, reflecting the change process from the initial state to the final state during the excavation process. The stress path is determined mainly by: analyzing geological data to understand the mechanical properties and stress conditions of the strata; using finite element software to simulate the stress distribution during the construction of the subway station; and combining on-site observations and historical data to determine the stress path.

[0058] Step 2: Based on the test results of the physical experiment, a numerical simulation model is selected, and the model parameters of the numerical simulation model are initialized. Then, the physical test parameters are input into the numerical simulation model to obtain the simulation results of the numerical simulation model: the vertical deformation of the soil and the subway station, and the simulated vertical deformation of the soil is obtained according to the vertical deformation of the soil and the subway station.

[0059] In this embodiment, the numerical simulation model is preferably the Cambridge model, which takes into account the elastic-plastic behavior, creep characteristics and relaxation effect of the soil. The model parameters include: elastic modulus, Poisson's ratio, plastic limit, plastic flow parameter, creep parameter, etc.

[0060] The model parameters are initialized using the test results of the physical test. For example, the elastic modulus and plastic limit are determined based on the stress-strain curve obtained from the triaxial compression test, the plastic flow parameters are determined based on the direct shear test results, and the creep parameters are determined based on the creep test results.

[0061] In the numerical simulation model, the stress path during the construction and operation of the subway station is set, and the simulation program is run to obtain the vertical deformation of the soil and the subway station. During the simulation process, the vertical deformation of each point in the soil is recorded, and statistical analysis is performed to obtain the simulated average vertical deformation of the soil or the simulated vertical deformation of the soil in a specific area.

[0062] The Cambridge model is an existing technology. When the Cambridge model is applied in this embodiment, it is necessary to adjust and optimize the model parameters according to the technical problems of this embodiment, so as to output the simulated vertical deformation of the soil.

[0063] Step 3: According to the simulation results of the numerical simulation model, excavation is carried out in the adjacent strata around the subway station to obtain the actual vertical deformation of the soil, and the vertical micro deformation of the subway station is obtained according to the actual vertical deformation of the soil.

[0064] According to the simulation results of the numerical simulation model, the area around the subway station with large vertical deformation is determined, excavation is carried out at the selected location to expose the soil surface, and the actual vertical deformation of the soil after excavation is measured using precision instruments (such as laser displacement sensors, GPS positioning systems, etc.). By comparing the actual vertical deformation of the soil with the simulation results and considering the interaction between the soil and the subway station, the vertical micro-deformation of the subway station is inferred.

[0065] Step 4: construct a correction model to correct the simulated vertical deformation of the soil and obtain the optimal vertical deformation of the soil simulation.

[0066] The correction model is preferably a deep learning model.

[0067] The simulated vertical deformation of the soil is corrected to obtain the optimal simulated vertical deformation of the soil, which specifically includes the following steps:

[0068] Step 41, obtaining a training set for the revised model.

[0069] Soil excavation cases involved in the construction of past subway stations are collected, and historical data in the soil excavation cases are preprocessed. The historical coordinate information of key positions of soil excavation, the historical stress path of the excavation process, the historical parameters of physical test tests, the historical parameters of soil characteristics, the historical parameters of adjacent strata structure, the historical vertical deformation of soil simulation, the historical model parameters of the numerical simulation model, and the historical optimal vertical deformation of soil simulation are obtained from the preprocessed historical data as a training set. In addition, the actual historical vertical deformation of the soil and the simulated historical vertical deformation of the soil in the historical data are imported into the data analysis software, and a scatter plot is drawn using the software, and a trend line is added. The correlation coefficient is calculated using the correlation coefficient formula or the built-in function of the software. If the correlation is significant, a regression analysis is performed, and the simulated historical vertical deformation of the soil that is close to the actual deformation is calibrated as the optimal historical vertical deformation of the simulated soil (the simulated historical vertical deformation of the soil whose error between the actual deformation and the simulated historical vertical deformation of the soil is within 0.05 is calibrated as the optimal historical vertical deformation of the simulated soil), and the optimal historical vertical deformation of the simulated soil is used as the output label. The input of the correction model is the actual vertical deformation of the soil and the simulated vertical deformation of the soil, and the output of the correction model is the optimal vertical deformation of the simulated soil.

[0070] The training set includes several sample sets, each of which includes: historical actual vertical deformation of soil, historical simulated vertical deformation of soil, historical model parameters of numerical simulation model, historical stress path of excavation process, historical coordinate information of key positions, historical parameters of physical test, historical parameters of soil characteristics, historical parameters of adjacent stratum structure, and historical optimal vertical deformation of soil simulation.

[0071] Step 42, constructing an iterative function according to the training set.

[0072] Specifically, this embodiment constructs an iterative function through the historical optimal vertical deformation of soil simulation and the optimal vertical deformation of soil simulation:

[0073] ;

[0074] ;

[0075] Where L represents the iterative function, represents the optimal vertical deformation of the soil simulation history of the i-th sample set, represents the optimal vertical deformation of soil simulation predicted by the modified model, n represents the maximum number of sample sets, M represents the modified model, D represents the optimized training set, represents the iteration condition, represents the threshold condition, F represents stopping the iteration operation, and T represents the iteration operation.

[0076] The goal of the above iterative function is to minimize and The difference between them is to minimize the gap between the historical optimal vertical deformation of soil simulation and the optimal vertical deformation of soil simulation. Through iteration, the model parameters of the modified model are gradually adjusted so that Closer , in order to improve the prediction accuracy.

[0077] Step 43, constructing a threshold condition for the iterative function.

[0078] The threshold condition is the condition for stopping the iteration, which determines a reasonable convergence standard, that is, how small the error in the iteration process can be considered to be close enough to the optimal solution. In this embodiment, the threshold condition is preferably ,when and The difference between them is less than 0.05, and the output That is the optimal vertical deformation of soil simulation, which can be used in the calculation of the correction coefficient.

[0079] Step 44, input the training set into the revised model for training, and obtain a trained revised model.

[0080] Step 45, inputting the actual vertical deformation of the soil and the simulated vertical deformation of the soil into the trained correction model, and outputting the first optimized vertical deformation of the simulated soil.

[0081] Step S46, calculating the first prediction error between the first optimized vertical deformation of the soil simulation and the training set.

[0082] Specifically, this step calculates the first prediction error It is obtained by calculating the difference between the first optimized vertical deformation and the simulated historical vertical deformation of the soil.

[0083] Step S47, compare the first prediction error with the threshold condition. If the first prediction error is less than or equal to the threshold condition, the output first optimized vertical deformation of soil simulation is the optimal vertical deformation of soil simulation; if the first prediction error is greater than the threshold condition, iterative processing is performed to obtain the optimal vertical deformation of soil simulation to realize the correction of the vertical deformation of soil simulation.

[0084] Specifically:

[0085] Step 471, re-perform physical test tests on the stress path during the excavation process of the key position to obtain the physical test test optimization parameters, optimize the model parameters after the initialization processing, input the physical test test optimization parameters into the numerical simulation model, and obtain the optimized vertical deformation of the soil and the subway station according to the numerical simulation model, and obtain the optimized simulated vertical deformation of the soil according to the optimized vertical deformation.

[0086] Optimized parameters are obtained through physical test to improve the prediction results of the modified model. Specifically, physical test is re-performed for the stress path during the excavation process at key locations to obtain physical test optimization parameters, which are used to optimize the numerical simulation model. The physical test optimization parameters are input into the numerical simulation model to optimize the numerical simulation model parameters. According to the optimized model parameters, the numerical simulation model is used to obtain the optimized vertical deformation of the soil and subway station. Based on the optimized vertical deformation, the optimized simulated vertical deformation of the soil is determined.

[0087] Step 472, adding the optimized soil simulated vertical deformation, the optimized model parameters, the physical test optimization parameters, and the optimized soil simulated vertical deformation as sub-sample sets to the training set to obtain an optimized training set.

[0088] Construct an optimized training set for training the revised model after iteration. In the process of adding the sub-sample set to the training set to obtain the optimized training set, it is preferred to perform data compression on the sub-sample set to ensure the generalization performance of the revised model.

[0089] Step S473, retrain the correction model using the optimized training set to obtain a trained optimized correction model.

[0090] Step S474, input the actual vertical deformation of the soil and the optimized simulated vertical deformation of the soil into the trained optimization correction model, and output the second optimized simulated vertical deformation of the soil.

[0091] Step S475, calculate the second prediction error between the second optimized vertical deformation of soil simulation and the optimized training set. If the second prediction error is less than or equal to the threshold condition, the second optimized vertical deformation of soil simulation is the optimal vertical deformation of soil simulation; if the second prediction error is greater than the threshold condition, return to execute steps 472-475 for iterative processing. When the second prediction error after iterative processing is less than or equal to the threshold condition or the number of iterations reaches the maximum number of iterations, stop the iteration. The second optimized vertical deformation of soil simulation after the last iterative processing is the optimal vertical deformation of soil simulation.

[0092] Whether further iteration is needed is determined by comparing the second prediction error with the threshold condition. Specifically, the second prediction error is obtained by comparing the difference between the second optimized vertical deformation and the historical vertical deformation of the soil simulation. When the second prediction error after iterative processing is less than or equal to the threshold condition, the iteration is stopped, or when the number of iterations reaches the maximum number of iterations set in advance, the iteration is stopped. The second optimized vertical deformation of the soil simulation after the last iterative processing is the optimal vertical deformation of the soil simulation. The maximum number of iterations in this embodiment is 50 times.

[0093] By comparing the actual deformation obtained from on-site excavation with the numerical simulation results, the deficiencies in the model can be discovered. The revised model can adjust the model parameters according to the difference between the actual deformation and the simulation results, making the model closer to the actual working conditions and improving the applicability and predictive ability of the model.

[0094] The simulated vertical deformation of soil refers to the vertical deformation of soil predicted by the numerical simulation model, which reflects the expected deformation of soil during excavation. However, due to the limitations of the numerical simulation model itself, such as model simplification, boundary condition assumptions, material parameter uncertainty and other factors, there is a deviation between the simulation results and the actual situation.

[0095] The actual vertical deformation of the soil refers to the actual vertical deformation of the soil obtained by on-site measurement means (such as displacement meters, strain gauges, etc.), which is an important basis for reflecting the real deformation of the soil. By comparing the difference between the actual vertical deformation and the simulated vertical deformation, problems in the numerical simulation can be found and the model can be corrected accordingly.

[0096] When we take the simulated vertical deformation of the soil and the actual vertical deformation of the soil as input, it is equivalent to establishing a feedback mechanism. Specifically, the difference between the simulated vertical deformation of the soil and the actual vertical deformation of the soil is regarded as the residual of the model, and then the residual is reduced by adjusting the model parameters, so that the simulation results are closer to the actual vertical deformation.

[0097] The advantage of the feedback mechanism is that it can continuously correct the model to make it gradually approach the real soil behavior. As the number of feedback increases, the accuracy of the model will gradually improve. Therefore, although the initial simulated vertical deformation of the soil may have a large deviation, the simulation results of the new numerical simulation model obtained by iterative updating can be used as a reference to gradually correct the model to ensure that the results output by the corrected model are getting closer and closer to the actual vertical deformation of the soil and reduce the error.

[0098] Step 5, obtaining a correction coefficient of the vertical micro-deformation of the subway station according to the simulated optimal vertical deformation of the soil and the actual vertical deformation of the soil;

[0099] The calculation formula of the correction coefficient is:

[0100] ;

[0101] Where N is the correction factor, It represents the actual vertical deformation of the soil. It represents the optimal vertical deformation of soil simulation output by the last iteration of the numerical simulation model.

[0102] This embodiment obtains accurate optimal vertical deformation of soil simulation through the correction model, and applies it to the calculation of correction coefficient of vertical micro-deformation of subway station, so as to obtain more accurate correction coefficient and improve the calculation accuracy for subsequent prediction value calculation.

[0103] Step 6: Obtain the predicted value of the vertical micro-deformation of the subway station according to the correction coefficient and the vertical micro-deformation of the subway station.

[0104] The calculation formula for the predicted value of vertical micro-deformation of subway stations is:

[0105] ;

[0106] In the formula, Y represents the predicted value of vertical micro-deformation of the subway station, N represents the correction coefficient, Indicates the vertical micro deformation of the subway station.

[0107] This embodiment uses a precise correction coefficient to correct the vertical micro-deformation of the subway station, which maximizes the calculation accuracy of the predicted value and provides better support for subsequent engineering decisions, especially in dealing with complex mechanical properties of soft soil and stratum heterogeneity. It reduces rework and additional costs caused by inaccurate predictions and can effectively control project costs.

[0108] See attached Figure 2 This embodiment also provides a vertical micro-deformation prediction system for a large subway station, which is used to execute the above-mentioned vertical micro-deformation prediction method for subway structures in subway double-sided extension, and includes the following modules:

[0109] Physical test module: select key locations around the subway station for sampling, and conduct physical test tests on the stress paths during excavation at key locations to obtain physical test results: physical test parameters;

[0110] Soil simulation vertical deformation acquisition module: connected with the physical test module, based on the physical test results, selects the numerical simulation model, initializes the model parameters of the numerical simulation model, and then inputs the physical test parameters into the numerical simulation model to obtain the simulation results of the numerical simulation model: the vertical deformation of the soil and the subway station, and obtains the simulated vertical deformation of the soil according to the vertical deformation of the soil and the subway station;

[0111] The actual vertical deformation acquisition module of the soil is connected to the simulated vertical deformation acquisition module of the soil, and is used to excavate the adjacent strata around the subway station according to the simulation results of the numerical simulation model to obtain the actual vertical deformation of the soil, and obtain the vertical micro deformation of the subway station according to the actual vertical deformation of the soil;

[0112] Correction model: connected with the soil simulation vertical deformation acquisition module, used to construct a correction model, correct the soil simulation vertical deformation, and obtain the optimal soil simulation vertical deformation;

[0113] Correction coefficient calculation module: connected with the correction model, used to obtain the correction coefficient according to the optimal vertical deformation of the soil simulation and the actual vertical deformation of the soil;

[0114] Prediction value calculation module: connected to the correction coefficient calculation module, used to obtain the prediction value of the vertical micro-deformation of the subway station according to the correction coefficient and the vertical micro-deformation of the subway station.

[0115] The above description is only a preferred specific implementation manner of the present application, and is not intended to limit the protection scope of the present application. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for predicting vertical micro-deformation of subway structures for subway double-sided extension, characterized by: The following steps are involved: Step 1, sampling is performed around the subway station, and a physical test is performed on the stress path during the excavation process to obtain the physical test results: physical test parameters; Step 2: Based on the test results of the physical experiment, a numerical simulation model is selected, and the model parameters of the numerical simulation model are initialized. Then, the test parameters of the physical experiment are input into the numerical simulation model to obtain the simulation results of the numerical simulation model: the vertical deformation of the soil and the subway station, and the simulated vertical deformation of the soil is obtained according to the vertical deformation of the soil and the subway station; Step 3: According to the simulation results of the numerical simulation model, excavation is performed in the adjacent strata around the subway station to obtain the actual vertical deformation of the soil, and the vertical micro deformation of the subway station is obtained according to the actual vertical deformation of the soil; Step 4, constructing a correction model to correct the simulated vertical deformation of the soil to obtain the optimal simulated vertical deformation of the soil; the input of the correction model is the actual vertical deformation of the soil and the simulated vertical deformation of the soil, and the output of the correction model is the optimal simulated vertical deformation of the soil; the correction model is a deep learning model; Step 5, obtaining a correction coefficient of the vertical micro-deformation of the subway station according to the simulated optimal vertical deformation of the soil and the actual vertical deformation of the soil; Step 6: Obtain the predicted value of the vertical micro-deformation of the subway station according to the correction coefficient and the vertical micro-deformation of the subway station.

2. The method for predicting vertical micro-deformation of subway structures in subway double-sided extension according to claim 1 is characterized by: Step 4, constructing a correction model, correcting the simulated vertical deformation of the soil, and obtaining the optimal simulated vertical deformation of the soil, specifically including the following steps: Step 41, obtaining a training set for the revised model; Step 42, constructing an iterative function according to the training set; Step 43, constructing a threshold condition for the iterative function; Step 44, input the training set into the correction model for training, and obtain a trained correction model; Step 45, inputting the actual vertical deformation of the soil and the simulated vertical deformation of the soil into the trained correction model, and outputting the first optimized vertical deformation of the simulated soil; Step S46, calculating the first prediction error between the first optimized vertical deformation of the soil simulation and the training set; Step S47, compare the first prediction error with the threshold condition. If the first prediction error is less than or equal to the threshold condition, the output first optimized vertical deformation of soil simulation is the optimal vertical deformation of soil simulation; if the first prediction error is greater than the threshold condition, iterative processing is performed to obtain the optimal vertical deformation of soil simulation.

3. The method for predicting vertical micro-deformation of subway structures in subway double-sided extension according to claim 2 is characterized by: The training set includes several sample sets, each of which includes: historical actual vertical deformation of soil, simulated historical vertical deformation of soil, simulated historical optimal vertical deformation of soil, historical model parameters of numerical simulation model, historical stress path of excavation process, historical coordinate information of sampling position, historical parameters of physical test, historical parameters of soil characteristics, and historical parameters of adjacent stratum structure.

4. The method for predicting vertical micro-deformation of subway structures in subway double-sided extension according to claim 2 is characterized by: If the first prediction error is greater than the threshold condition, an iterative process is performed to obtain the optimal vertical deformation of the soil simulation, which specifically includes the following steps: Step 471, re-perform physical test on the stress path during the excavation process of the key position to obtain physical test optimization parameters, optimize the model parameters after the initialization processing, input the physical test optimization parameters into the numerical simulation model, and obtain the optimized vertical deformation of the soil and the subway station according to the numerical simulation model, and obtain the optimized simulated vertical deformation of the soil according to the optimized vertical deformation; Step 472, adding the optimized simulated vertical deformation of the soil, the optimized model parameters, the physical test optimization parameters, and the optimized simulated vertical deformation of the soil as sub-sample sets to the training set to obtain an optimized training set; Step S473, retraining the correction model using the optimized training set to obtain a trained optimized correction model; Step S474, inputting the actual vertical deformation of the soil and the optimized simulated vertical deformation of the soil into the trained optimization correction model, and outputting the simulated second optimized vertical deformation of the soil; Step S475, calculate the second prediction error between the second optimized vertical deformation of soil simulation and the optimized training set. If the second prediction error is less than or equal to the threshold condition, the second optimized vertical deformation of soil simulation is the optimal vertical deformation of soil simulation; if the second prediction error is greater than the threshold condition, return to execute steps 472-475 for iterative processing. When the second prediction error after iterative processing is less than or equal to the threshold condition or the number of iterations reaches the maximum number of iterations, stop the iteration. The second optimized vertical deformation of soil simulation after the last iterative processing is the optimal vertical deformation of soil simulation.

5. The method for predicting vertical micro-deformation of subway structures in subway double-sided extension according to claim 2 is characterized by: The iterative function is: ； ； Where L represents the iterative function, represents the optimal vertical deformation of the soil simulation history of the i-th sample set, represents the optimal vertical deformation of soil simulation predicted by the modified model, n represents the maximum number of sample sets, M represents the modified model, D represents the optimized training set, represents the iteration condition, represents the threshold condition, F represents stopping the iteration operation, and T represents the iteration operation.

6. The method for predicting vertical micro-deformation of subway structures in subway double-sided extension according to claim 3 is characterized by: The training set is obtained from historical soil excavation cases, and the simulated historical optimal vertical deformation of the soil is used as an output label of the correction model.

7. The method for predicting vertical micro-deformation of subway structure in subway double-sided extension according to claim 1 is characterized by: The calculation formula of the correction coefficient is: ； Where N is the correction factor, Indicates the actual vertical deformation of the soil. It represents the optimal vertical deformation of soil simulation output by the last iteration of the numerical simulation model.

8. The method for predicting vertical micro-deformation of subway structures in subway double-sided extension according to claim 1 is characterized by: The calculation formula for the predicted value of vertical micro-deformation of subway stations is: ； In the formula, Y represents the predicted value of vertical micro-deformation of the subway station, N represents the correction coefficient, Indicates the vertical micro deformation of the subway station.

9. A vertical micro-deformation prediction system for a large subway station, used to implement the vertical micro-deformation prediction method for a subway structure in a subway double-sided extension as described in any one of claims 1 to 8, characterized in that: Includes the following modules: Physical test module: sampling around the subway station, and conducting physical test on the stress path during the excavation process to obtain the physical test results: physical test parameters; Soil simulation vertical deformation acquisition module: connected with the physical test module, based on the physical test results, selects the numerical simulation model, initializes the model parameters of the numerical simulation model, and then inputs the physical test parameters into the numerical simulation model to obtain the simulation results of the numerical simulation model: the vertical deformation of the soil and the subway station, and obtains the simulated vertical deformation of the soil according to the vertical deformation of the soil and the subway station; The actual vertical deformation acquisition module of the soil is connected to the simulated vertical deformation acquisition module of the soil, and is used to excavate the adjacent strata around the subway station according to the simulation results of the numerical simulation model to obtain the actual vertical deformation of the soil, and obtain the vertical micro deformation of the subway station according to the actual vertical deformation of the soil; Correction model: connected with the soil simulation vertical deformation acquisition module, used to construct a correction model, correct the soil simulation vertical deformation, and obtain the optimal soil simulation vertical deformation; Correction coefficient calculation module: connected with the correction model, used to obtain the correction coefficient according to the optimal vertical deformation of the soil simulation and the actual vertical deformation of the soil; Prediction value calculation module: connected to the correction coefficient calculation module, used to obtain the prediction value of the vertical micro-deformation of the subway station according to the correction coefficient and the vertical micro-deformation of the subway station.

Images