Pipe roofing construction ground surface settlement prediction method combined with multi-factor influence analysis

Abstract

The invention discloses a pipe-roofing construction ground surface settlement prediction method combined with multi-factor influence analysis, and relates to the technical field of underground engineering settlement prediction, and the method comprises the steps: carrying out the simulation prediction of the pipe-roofing jacking influence, the pipe-roofing structure deflection deformation and the soil arching effect; generating predicted jacking influence data, predicted deflection deformation data and predicted soil arching effect data; performing pipe roofing construction ground surface settlement prediction to generate a static ground surface settlement prediction result; and performing ground surface settlement dynamic compensation analysis in combination with the regional train operation characteristics and the regional soil mass attribute information, determining a dynamic compensation increment, adjusting a static ground surface settlement prediction result, and outputting the ground surface settlement prediction result. The technical problem that the accuracy of the prediction result is insufficient due to the fact that multiple factors are not fully considered in the existing pipe roofing construction ground surface settlement prediction is solved, the purpose of accurately predicting the pipe roofing construction ground surface settlement by integrating the multiple factors is achieved, and the technical effect of improving the prediction accuracy of the pipe roofing construction ground surface settlement is achieved.

Description

Technical Field

[0001] The invention relates to the technical field of underground engineering settlement prediction, and in particular to a method for predicting surface settlement during pipe curtain construction in combination with multi-factor influence analysis. Background Art

[0002] During the pipe-curtain construction process, the prediction of surface settlement is extremely important. Traditional methods for predicting surface settlement during pipe-curtain construction have many shortcomings. They fail to comprehensively consider multiple influencing factors such as regional soil property information, steel pipe property information, expected jacking construction parameters, and regional train operation characteristics, resulting in inaccurate simulation predictions of the impact of pipe-curtain jacking, pipe-curtain structure deflection and deformation, and soil arching effect. At the same time, in the construction of surface settlement prediction models, there is a lack of effective integrated learning methods, making it difficult to accurately integrate various types of prediction data. These problems make the existing methods inaccurate in predicting surface settlement during pipe-curtain construction and unable to meet the actual needs of the project.

[0003] The existing prediction of surface settlement during pipe-roof construction has a technical problem of insufficient accuracy of prediction results due to the failure to fully consider multiple factors. Summary of the invention

[0004] The present application provides a method for predicting surface settlement during pipe-roof construction in combination with multi-factor impact analysis, which is used to solve the technical problem that the prediction results are not accurate enough due to insufficient consideration of multiple factors in the existing prediction of surface settlement during pipe-roof construction.

[0005] In view of the above problems, the present application provides a method for predicting surface settlement during pipe-roof construction by combining multi-factor impact analysis, the method comprising: Combined with the regional soil property information, steel pipe property information and expected jacking construction parameters, simulation predictions of the pipe curtain jacking impact, pipe curtain structure deflection deformation and soil arching effect are carried out respectively to generate predicted jacking impact data, predicted deflection deformation data and predicted soil arching effect data; based on the principle of integrated learning, the surface settlement of pipe curtain construction is predicted according to the predicted jacking impact data, predicted deflection deformation data and predicted soil arching effect data to generate static surface settlement prediction results; combined with the regional train operation characteristics and the regional soil property information, dynamic compensation analysis of surface settlement is carried out to determine the dynamic compensation increment, adjust the static surface settlement prediction results, and output the surface settlement prediction results.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: Combined with the regional soil attribute information, steel pipe attribute information and expected jacking construction parameters, simulation predictions of pipe curtain jacking influence, pipe curtain structure deflection deformation and soil arching effect are carried out respectively, and predicted jacking influence data, predicted deflection deformation data and predicted soil arching effect data are generated; based on the principle of integrated learning, the surface settlement of pipe curtain construction is predicted according to the predicted jacking influence data, predicted deflection deformation data and predicted soil arching effect data, and static surface settlement prediction results are generated; dynamic compensation analysis of surface settlement is carried out, dynamic compensation increment is determined, the static surface settlement prediction results are adjusted, and surface settlement prediction results are output. The purpose of realizing accurate prediction of surface settlement of pipe curtain construction by integrating multiple factors is achieved, and the technical effect of improving the accuracy of surface settlement prediction of pipe curtain construction is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0008] Figure 1 A schematic diagram of a flow chart of a method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis provided in an embodiment of the present application; Figure 2 A schematic flow chart of establishing a soil-pipe-roof interaction model for a method for predicting surface settlement during pipe-roof construction that combines multi-factor impact analysis provided in an embodiment of the present application. DETAILED DESCRIPTION

[0009] The present application provides a method for predicting surface settlement during pipe curtain construction that combines multi-factor impact analysis, in order to solve the technical problem that the prediction results are not accurate enough due to insufficient consideration of multiple factors in the existing prediction of surface settlement during pipe curtain construction.

[0010] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0011] Examples, such as Figure 1 As shown, the present application provides a method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis, the method comprising: Step S100: combining regional soil property information, steel pipe property information and expected jacking construction parameters, respectively simulate and predict the impact of pipe curtain jacking, pipe curtain structure deflection and deformation, and soil arching effect, and generate predicted jacking impact data, predicted deflection and deformation data, and predicted soil arching effect data.

[0012] Specifically, we first need to collect information on the soil type, soil layer structure, soil mechanical properties, and groundwater level of the construction area, as well as information on the steel pipe properties such as material, strength, diameter, and wall thickness of the steel pipe used for pipe curtain construction, as well as expected jacking speed and expected jacking sequence of the steel pipe jacking operation. Finite element analysis is used to establish a soil-pipe curtain interaction model based on the collected information. Based on this model, the pipe curtain jacking process is simulated according to the expected jacking construction parameters. During the simulation process, the data on the impact of the pipe curtain jacking on the surrounding soil, such as soil displacement and stress change, can be obtained respectively, which constitute the predicted jacking impact data; at the same time, the deflection deformation data of the pipe curtain structure during the jacking process is obtained to reflect the deformation of the pipe curtain itself; in addition, the data on the soil arch effect generated by the soil under the action of the pipe curtain construction can be predicted, including the predicted arch pressure distribution. These three types of data are fully simulated and generated, providing key basic information for subsequent surface settlement prediction.

[0013] Step S200: Based on the principle of integrated learning, the surface settlement of pipe-roof construction is predicted according to the predicted jacking influence data, the predicted deflection deformation data and the predicted soil arching effect data, and a static surface settlement prediction result is generated.

[0014] Specifically, firstly, based on the principle of ensemble learning, Q (Q is an integer greater than or equal to 2) surface settlement prediction branches are constructed by using an ensemble learning operator that includes at least a feedforward neural network and a random decision forest. Sample jacking influence data sets, sample steel pipe deflection deformation data sets, and sample soil arching effect data sets are collected from the historical pipe curtain construction records. At the same time, the corresponding sample surface settlement data sets, including settlement amount, settlement rate, and settlement distribution, are obtained to construct a sample surface settlement data set. These sample data are divided into Q parts, and then divided into Q training sets and Q validation sets in proportion. These training sets and validation sets are used to perform supervised training and validation training on the Q surface settlement prediction branches respectively. During the training process, the model parameters are continuously adjusted until each branch meets the expected conditions, thereby obtaining Q converged surface settlement prediction branches. Finally, based on the weighted integration principle, the performance and importance of each branch are comprehensively considered, different weights are assigned to each convergent branch, and these Q convergent surface settlement prediction branches are combined to construct a surface settlement prediction model. The generated predicted jacking impact data, predicted deflection deformation data and predicted soil arch effect data are input into the model, and the model outputs the static surface settlement prediction results of pipe curtain construction, laying the foundation for subsequent dynamic compensation analysis and final surface settlement prediction.

[0015] Step S300: Performing dynamic compensation analysis on surface settlement in combination with regional train operation characteristics and regional soil property information, determining a dynamic compensation increment, adjusting the static surface settlement prediction result, and outputting the surface settlement prediction result.

[0016] Specifically, the regional train operation characteristics are obtained, including information such as historical passing speed, historical load average and expected number of passes. These data can be collected from the railway operation management system or related monitoring equipment. At the same time, combined with the regional soil property information collected previously, the soil-train dynamic interaction model is constructed using the finite element analysis method. In this model, the dynamic load effect on the soil caused by the train running is simulated, and the influence of factors such as vibration and impact caused by train running on the stress-strain state of the soil is considered. Then, the soil-train dynamic interaction model is combined with the previously established soil-pipe curtain interaction model to conduct a dynamic compensation analysis of surface settlement. Through model calculation, the dynamic compensation increment is obtained, including the settlement increment and the settlement rate increment. These increments reflect the additional impact of train operation on surface settlement. Finally, the dynamic compensation increment is applied to the static surface settlement prediction results to adjust them. By adding the settlement increment to the static predicted settlement, the settlement change trend over time is adjusted according to the settlement rate increment, so as to obtain and output a surface settlement prediction result that is more in line with the actual situation. This result comprehensively considers the impact of pipe curtain construction and train operation on surface settlement, thereby improving the accuracy of the prediction.

[0017] In a possible implementation, step S100 further includes: Step S110: collecting soil type, soil layer structure, soil mechanical properties and groundwater level of the area to be constructed as soil attribute information of the area.

[0018] Step S120: collecting the material, strength, diameter and wall thickness of the steel pipe used for the pipe curtain construction as the steel pipe attribute information.

[0019] Step S130: collecting the expected jacking speed and expected jacking sequence of the steel pipe jacking operation as the expected jacking construction parameters.

[0020] Specifically, in order to obtain regional soil property information, it is necessary to conduct a comprehensive geological survey of the area to be constructed, and adopt a variety of means, such as geological drilling, to collect soil samples from different depths and accurately analyze the type of soil and soil layer structure. Through particle analysis tests, determine whether the soil is sandy, clay, or silty soil; with the help of geophysical exploration technologies such as geological radar, the distribution and changes of soil layers can be more intuitively understood, and the soil layer structure can be determined. For soil mechanical properties, direct shear tests and triaxial compression tests will be carried out to determine key parameters such as shear strength and elastic modulus of the soil. At the same time, groundwater level monitoring wells are set up at the construction site. By regularly observing water level changes, groundwater level information can be accurately obtained. These data comprehensively constitute the regional soil property information, which can provide a basic basis for subsequent simulation predictions in terms of soil.

[0021] For steel pipes used in pipe curtain construction, check the quality certification documents, product manuals and other materials attached to the steel pipes. These documents usually clearly mark the material information of the steel pipes, such as common steel models such as Q345B and X65. They also include the strength parameters of the steel pipes, such as key indicators such as yield strength and tensile strength. They directly reflect the mechanical properties of the steel pipes and are of great significance for judging the bearing capacity of the steel pipes during the pipe curtain construction process. In addition to consulting the documents, in order to ensure the accuracy and reliability of the data, measuring tools are also used to measure the diameter and wall thickness of the steel pipes on the spot. When measuring the pipe diameter, the commonly used tools are high-precision vernier calipers or outside micrometers. At different positions of the steel pipe, such as the two ends and the middle part, multiple measurements are taken to obtain the average value to obtain more accurate pipe diameter data; for wall thickness measurement, with the help of an ultrasonic thickness gauge, multiple measurement points are evenly selected on the surface of the steel pipe for measurement, so as to obtain reliable wall thickness values. The information on steel pipe material, strength, diameter and wall thickness collected through the above operations together constitute complete steel pipe property information, providing indispensable data support for subsequent surface settlement prediction of pipe curtain construction in combination with other factors.

[0022] The expected jacking speed and expected jacking sequence of the steel pipe jacking operation are collected as key expected jacking construction parameters. The construction team needs to determine these parameters based on the detailed construction design plan, taking into account the geological conditions on site, the design requirements of the pipe curtain, and the performance of the jacking equipment. For the expected jacking speed, it is necessary to balance the stability of the soil and the construction efficiency. If the speed is too fast, it may cause excessive disturbance of the soil and cause large surface settlement; if the speed is too slow, it will affect the construction progress. By analyzing the mechanical properties of the soil at the construction site and combining the experience of previous similar projects, a reasonable jacking speed value is determined, such as how many meters per hour, and it is recorded. The determination of the expected jacking sequence needs to be based on the design layout of the pipe curtain to ensure that the jacking of each steel pipe will not have an adverse effect on the steel pipes that have been jacked, and avoid mutual interference. For example, jacking from one side to the other side in sequence, or staggered jacking at specific intervals, etc. The determined expected jacking sequence is recorded in detail to clarify the jacking sequence of each steel pipe. These accurately collected expected jacking speeds and expected jacking sequences provide important construction parameter basis for the subsequent simulation prediction of pipe curtain construction based on regional soil property information and steel pipe property information, and play a key role in improving the accuracy of surface settlement prediction.

[0023] In a possible implementation, step S100 further includes: Step S140: Based on finite element analysis, a soil-pipe-roof interaction model is established according to the regional soil attribute information and the steel pipe attribute information.

[0024] Step S150: Utilizing the soil-pipe curtain interaction model, perform pipe curtain jacking simulation according to the expected jacking construction parameters, and output predicted jacking impact data and predicted deflection deformation data, wherein the predicted jacking impact data includes first surface settlement prediction information, and the first surface settlement prediction information includes a first predicted settlement amount, a first predicted settlement rate, and a first predicted settlement distribution.

[0025] Specifically, firstly, a soil-pipe curtain interaction model is established based on finite element analysis, and geometric modeling algorithms, meshing algorithms and finite element equation solving algorithms are applied. Firstly, a geometric model is constructed based on the regional soil attribute information and steel pipe attribute information using the geometric modeling algorithm. For the soil, according to its soil layer structure information, a hierarchical modeling algorithm is used to treat different soil layers as independent geometric entities and combine them according to the actual soil layer distribution; for the steel pipe, according to the pipe diameter and wall thickness information, a parametric modeling algorithm is used to generate the pipe curtain geometric model by inputting parameters such as pipe diameter and wall thickness. Then, the meshing algorithm is used to process the constructed geometric model. For example, the Delaunay triangulation algorithm is used to perform two-dimensional or three-dimensional meshing on the soil and pipe curtain geometric models. This algorithm discretizes the continuous model by constructing non-overlapping triangular (two-dimensional) or tetrahedral (three-dimensional) mesh units on the surface or space of the geometric model. In the process of meshing, the density of the mesh is adjusted according to the geometric characteristics of the model and the expected calculation accuracy, and a finer mesh is set in key parts (such as the contact area between the pipe curtain and the soil) to improve the calculation accuracy. Then, the soil-pipe-roof interaction model is established based on the finite element equation solving algorithm. According to the material properties of the soil and steel pipe (such as elastic modulus, Poisson's ratio, etc.), the material constitutive relationship is determined and substituted into the finite element equation. The finite element equation is usually derived based on the virtual work principle or variational principle, which describes the relationship between the displacement of each node in the model and the load. By solving these equations, the mechanical response of the soil and the pipe-roof under interaction is obtained, and the soil-pipe-roof interaction model is established.

[0026] The established soil-pipe curtain interaction model is used to simulate the pipe curtain jacking. The explicit dynamics algorithm is used to simulate the jacking process of the pipe curtain in the model according to the expected jacking construction parameters (such as the expected jacking speed and sequence). The algorithm considers the contact and friction between the pipe curtain and the soil by gradually calculating the change in the mechanical state of the pipe curtain and the soil in each time step. In each time step, the geometric configuration of the model is updated according to the displacement and speed of the pipe curtain, and the stress and strain distribution of the soil and the pipe curtain are recalculated. As the simulation proceeds, the predicted jacking influence data and predicted deflection deformation data are obtained through the data extraction algorithm. For the first surface settlement prediction information in the predicted jacking influence data, the displacement monitoring algorithm is used to select the key nodes on the surface in the model, and the vertical displacement of these nodes during the pipe curtain jacking process is monitored in real time, so as to obtain the first predicted settlement; the first predicted settlement rate is obtained by calculating the ratio of the change in the displacement of the nodes in adjacent time steps to the time interval; the settlement data of different surface nodes are sorted and visualized to obtain the first predicted settlement distribution. For predicting the deflection deformation data, monitoring points are set on the pipe curtain model to monitor the deformation of the pipe curtain during the jacking process and obtain the deflection deformation data of the pipe curtain.

[0027] In one possible implementation, Figure 2 As shown, step S140 also includes: Step S141: geometric feature modeling is performed according to the regional soil attribute information and the steel pipe attribute information to generate a soil geometry model and a pipe curtain geometry model.

[0028] Step S142: meshing the body geometric model and the pipe-roof geometric model according to a predetermined size to obtain a soil body mesh model and a pipe-roof mesh model.

[0029] Step S143: Based on finite element analysis, a soil-pipe-roof interaction model is established according to the soil grid model and the pipe-roof grid model.

[0030] Specifically, geometric feature modeling is carried out based on the regional soil attribute information and steel pipe attribute information collected in the early stage. For the soil part, professional modeling software (such as modeling tools provided by AutoCAD, ABAQUS, etc.) is used to construct it according to information such as soil type and soil layer structure. For example, if there are multiple layers of soil with different properties in the area, geometric figures representing different soil layers are drawn in sequence according to the actual thickness and distribution range of each soil layer, and these figures are combined to form a complete soil geometric model, thereby accurately presenting the spatial structural characteristics of the soil. For the construction of the pipe curtain geometric model, according to the attribute parameters such as the diameter and wall thickness of the steel pipe, these data are input into the modeling software, and the parametric modeling function of the software is used to generate steel pipe geometric figures that meet the actual size, and these steel pipe figures are arranged and combined according to the layout of the pipe curtain design to generate the pipe curtain geometric model.

[0031] After completing the construction of the geometric model, it is necessary to mesh the soil geometric model and the pipe curtain geometric model according to the predetermined size, so as to discretize the continuous model and facilitate subsequent numerical calculations. Select a suitable meshing tool (such as the meshing module in ABAQUS) in the modeling software, and divide the soil geometric model according to the pre-set mesh size parameters. During the division process, for different parts of the soil, the mesh density will be reasonably adjusted according to the complexity of the force and the requirements of calculation accuracy. For example, in the area where the pipe curtain contacts the soil, a finer mesh will be set due to the complex stress changes; in the soil area far away from the pipe curtain, the mesh can be appropriately sparse to balance the calculation accuracy and efficiency. Similarly, the pipe curtain geometric model is also meshed according to the predetermined size to ensure that each part of the pipe curtain can be accurately discretized, and finally the soil mesh model and the pipe curtain mesh model are obtained respectively.

[0032] The obtained soil mesh model and pipe curtain mesh model are imported into professional finite element analysis software, such as ANSYS, ABAQUS, etc. In the software environment, the material parameters of the soil and steel pipe are set in detail according to the regional soil attribute information and steel pipe attribute information. For the soil, input parameters such as elastic modulus, Poisson's ratio, density, etc. These parameters determine the deformation characteristics and stress distribution of the soil when subjected to force; for the steel pipe, set its elastic modulus, yield strength, density and other parameters to accurately describe the mechanical properties of the steel pipe. Next, define the interaction relationship between the soil and the pipe curtain, select the contact algorithm in the software, such as setting it to friction contact, and determine the friction coefficient according to engineering experience or relevant test data, so as to simulate the mechanical behavior of the soil and the pipe curtain when they contact each other in actual construction, including the force transmission and relative displacement between the two. Then, apply boundary conditions, and constrain the boundaries of the soil model according to the actual construction scene, such as fixing the bottom boundary of the soil to limit its displacement in all directions; apply initial conditions that meet the actual jacking situation to the pipe curtain model, such as presetting the jacking direction and initial velocity. After completing the above settings, the finite element analysis software will discretize the entire soil-pipe curtain system into a finite number of units based on the finite element theory, and perform mechanical analysis on each unit. By solving a series of complex linear equations, the mechanical response between the soil and the pipe curtain under various working conditions is simulated, including stress, strain distribution, and displacement changes. Finally, a model that can truly reflect the interaction between the soil and the pipe curtain is successfully established, which provides a solid foundation for the subsequent use of the model to simulate the pipe curtain jacking process, predict the impact of pipe curtain jacking, pipe curtain structure deflection deformation, and analyze soil arch effect, etc., which helps to more accurately predict the surface settlement caused by pipe curtain construction.

[0033] In a possible implementation, step S150 further includes: Step S151: using the soil-pipe-roof interaction model, performing pipe-roof jacking simulation according to the expected jacking construction parameters, predicting soil deformation, and outputting predicted soil arching effect data, wherein the predicted soil arching effect data includes predicted arching pressure distribution.

[0034] Specifically, the simulation of pipe-roof jacking is carried out by using the constructed soil-pipe-roof interaction model and the expected jacking construction parameters collected, including the expected jacking speed and the expected jacking sequence. The expected jacking construction parameters are input into the finite element analysis software, and the software simulates the process of pipe-roof jacking one by one based on these parameters. When simulating the jacking of the pipe-roof, the software calculates the mechanical response of the soil and pipe-roof at each time step in real time according to the soil and pipe-roof material properties, contact relationships and boundary conditions preset in the model. As the pipe-roof is jacked, the soil is squeezed and disturbed, and its internal stress field and displacement field are constantly changing. In this process, the software focuses on the deformation of the soil to predict the soil arching effect. The soil arching effect is an important phenomenon in the mechanical behavior of the soil in the pipe-roof construction, which has a key influence on the surface settlement and soil stability. The software simulates the arched structure formed inside the soil by calculating and analyzing the deformation of the soil, and then obtains the predicted soil arching effect data. Among them, the predicted arch pressure distribution is the key output data, which shows the pressure size and distribution at different positions on the arch structure formed inside the soil under the influence of the pipe curtain jacking. By analyzing these predicted arch pressure distribution data, we can understand the stability, bearing capacity and impact range of the soil arch on the surrounding soil, which provides an important basis for the subsequent accurate prediction of surface settlement during pipe curtain construction, and also helps to evaluate the stability and safety of the soil during the construction process.

[0035] In a possible implementation, step S200 further includes: Step S210: construct Q surface subsidence prediction branches according to Q integrated learning operators, wherein the integrated learning operator includes at least a feedforward neural network and a random decision forest, and Q is an integer greater than or equal to 2.

[0036] Step S220: According to the historical pipe-roof construction records, sample data are collected to perform supervised training and verification training on the Q surface settlement prediction branches respectively, and a surface settlement prediction model is constructed in an integrated manner.

[0037] Step S230: inputting the predicted jacking impact data, predicted deflection deformation data and predicted soil arching effect data into the surface settlement prediction model, and outputting a static surface settlement prediction result.

[0038] Specifically, in order to improve the accuracy and reliability of prediction, the strategy of integrated learning is used to construct Q surface settlement prediction branches with the help of Q integrated learning operators. Among them, the integrated learning operator covers at least feedforward neural network and random decision forest, and Q is an integer greater than or equal to 2. First, the specific value of Q is determined based on the actual needs of the project and the characteristics of the data. Then, professional tools and frameworks in the field of machine learning are used, such as TensorFlow and PyTorch libraries based on Python to build feedforward neural networks, and random decision forests are built with the help of Scikit-learn libraries. For feedforward neural networks, by designing the appropriate number of network layers, the number of neurons in each layer, and the activation function, it has a strong nonlinear mapping ability and can effectively mine the complex features in the data. The random decision forest generates multiple decision trees, randomly selects features and samples during the construction process, and finally improves the generalization performance and stability of the model by integrating the prediction results of multiple decision trees. Through these operations, Q surface settlement prediction branches with different structural and functional characteristics were successfully constructed. These branches will study and analyze the surface settlement related data of pipe curtain construction from different angles, providing a basis for the subsequent construction of a comprehensive surface settlement prediction model.

[0039] The sample data are collected using historical pipe curtain construction records. These records contain rich construction information, such as regional soil properties, steel pipe properties, jacking construction parameters and corresponding actual surface settlement data under different construction conditions. Sample jacking influence data sets, sample steel pipe deflection deformation data sets and sample soil arching effect data sets are extracted from these records. At the same time, the corresponding sample surface settlement data sets are obtained, covering settlement amount, settlement rate and settlement distribution, and then the sample surface settlement data sets are constructed. Subsequently, these sample data are divided into Q parts, and divided into Q training sets and Q validation sets according to a certain ratio. Using these Q training sets and Q validation sets, supervised training and validation training are performed on the constructed Q surface settlement prediction branches respectively. During the training process, the parameters of the model are continuously adjusted, such as the weight and bias of the neural network, the number of trees of the random decision forest, the maximum depth, etc., so that the model gradually learns the rules in the data until each branch meets the preset expected conditions, such as the accuracy rate and loss value reach a certain standard, thereby obtaining Q converged surface settlement prediction branches. Finally, based on the weighted integration principle, different weights are assigned to each branch according to its performance in the training and validation process, and the Q converged surface subsidence prediction branches are combined to construct a comprehensive surface subsidence prediction model.

[0040] After the model is built, the predicted jacking impact data, predicted deflection deformation data, and predicted soil arching effect data obtained in the simulation prediction are input into the constructed surface settlement prediction model. After receiving these data, the model conducts a comprehensive analysis and processing of the data according to its internal algorithm and the weight relationship obtained through training, and finally outputs the static surface settlement prediction results of pipe curtain construction. This result provides basic data for the subsequent consideration of the impact of dynamic factors such as regional train operation on surface settlement, which helps to further improve the accuracy of surface settlement prediction.

[0041] In a possible implementation, step S220 further includes: Step S221: According to the historical pipe curtain construction records, collect sample jacking influence data sets, sample steel pipe deflection deformation data sets and sample soil arch effect data sets, and obtain sample surface settlement data under different sample jacking influence data, sample steel pipe deflection deformation data and sample soil arch effect data, and construct a sample surface settlement data set, wherein the surface settlement data includes settlement amount, settlement rate and settlement distribution.

[0042] Step S222: taking the sample jacking influence data set, the sample steel pipe deflection deformation data set, the sample soil arching effect data set and the sample surface settlement data set as sample data and dividing them into Q parts equally, and obtaining Q training sets and Q validation sets after dividing them in proportion.

[0043] Step S223: Using the Q training sets and the Q validation sets, respectively, the Q surface subsidence prediction branches are supervised and validated until the expected conditions are met, thereby obtaining Q converged surface subsidence prediction branches.

[0044] Step S224: Based on the weighted integration principle, the Q convergent surface subsidence prediction branches are combined to obtain the surface subsidence prediction model.

[0045] Specifically, we deeply excavate the key data in the historical pipe curtain construction records, and screen out various types of data closely related to the current research by consulting the detailed archives of multiple previous pipe curtain construction projects. For the sample jacking impact data set, the impact data of steel pipe jacking on the surrounding soil under different construction conditions, such as soil stress changes and displacement, will be extracted; the sample steel pipe deflection deformation data set focuses on the deformation of the steel pipe itself during the construction process, including the deflection values ​​at different positions and time points; the sample soil arch effect data set collects relevant data when the soil forms a soil arch due to pipe curtain construction, such as the shape parameters of the soil arch, pressure distribution characteristics, etc. At the same time, for these different sample jacking impact data, sample steel pipe deflection deformation data, and sample soil arch effect data, the corresponding sample surface settlement data are accurately obtained using the data in the field monitoring equipment records and engineering reports. These surface settlement data cover settlement, that is, the displacement of the surface in the vertical direction during the construction process; settlement rate, which reflects the speed of surface settlement changes over time; settlement distribution, which shows the difference in settlement at different locations on the surface. The collected data are sorted, classified and integrated to construct a sample surface settlement dataset, which comprehensively and meticulously reflects the relationship between different construction factors and surface settlement, providing a crucial data basis for subsequent training and optimization of the surface settlement prediction model.

[0046] The constructed sample jacking effect data set, sample steel pipe deflection deformation data set, sample soil arch effect data set and sample surface settlement data set are integrated together and regarded as complete sample data. Since the Q surface settlement prediction branches need to be trained and verified later, in order to ensure that each branch can obtain sufficient and balanced data learning, these sample data are divided into Q parts. When dividing, a random sampling method is used to ensure that each data can cover samples of various construction conditions and settlement conditions to avoid data bias. After the equal division is completed, the Q parts of data are selected according to a certain proportion. Usually, a large proportion (such as 70%-80%) of the data is set as the training set for training the Q surface settlement prediction branches, so that the model can learn the laws and characteristics in the data; the remaining proportion of the data is used as the verification set to test the generalization ability of the model during the training process, ensuring that the model can maintain good prediction performance when facing new data. Through such proportional division, Q training sets and Q verification sets are finally obtained, which provides reasonable data support for subsequent model training and verification work.

[0047] The Q constructed surface subsidence prediction branches are supervised and verified by using the divided Q training sets and Q validation sets. In the supervised training stage, each prediction branch takes the sample data in the training set as input, and continuously adjusts the parameters of the model (such as the weight of the feedforward neural network, the number of trees in the random decision forest, etc.) to make the prediction results of the model as close as possible to the true value in the sample surface subsidence data. At the same time, during the training process, the model is verified using the validation set to check the performance of the model on unseen data. The training and verification process is repeated until each prediction branch meets the preset expected conditions, such as the loss function converges to a certain range, the accuracy reaches a specific standard, etc. At this time, Q converged surface subsidence prediction branches are obtained, and each branch has a certain prediction ability.

[0048] Based on the weighted integration principle, different weights are assigned to the Q convergent surface settlement prediction branches, taking into account the differences in learning data characteristics and prediction accuracy of different prediction branches. The weight allocation is usually determined based on the performance of each branch in the verification phase, and branches with better performance are given higher weights. These branches with different weights are combined so that their advantages complement each other, and finally a comprehensive surface settlement prediction model is constructed. This model can more accurately handle the complex pipe curtain construction surface settlement prediction problem and provide strong support for the subsequent prediction of static surface settlement results.

[0049] In a possible implementation, step S300 further includes: Step S310: Acquire regional train operation characteristics, wherein the regional train operation characteristics include historical passing speed, historical load average and expected passing times.

[0050] Step S320: Based on finite element analysis, a soil-train dynamic interaction model is constructed according to the regional train operation characteristics and regional soil property information.

[0051] Step S330: Performing surface settlement dynamic compensation analysis using the soil-pipe-roof interaction model and the soil-train dynamic interaction model to obtain a dynamic compensation increment, wherein the dynamic compensation increment includes a settlement increment and a settlement rate increment.

[0052] Specifically, in order to comprehensively consider the impact of regional train operation on the surface settlement of pipe curtain construction, it is necessary to obtain regional train operation characteristics and obtain relevant data from the railway operation management department. These data include historical passing speed, historical load average and expected number of passes. The historical passing speed and load average reflect the actual situation of past train operation, which can help analyze the influence of long-term train operation on soil; the expected number of passes is combined with railway planning information to make the prediction more forward-looking. By collecting and organizing these data, basic information on train operation can be provided for subsequent analysis.

[0053] Based on the finite element analysis method, the soil-train dynamic interaction model is constructed by combining the regional train operation characteristics and regional soil property information. In the finite element analysis software (such as ANSYS, ABAQUS, etc.), the regional soil property information, including soil type, soil layer structure, soil mechanical properties, etc., is first input to establish the initial model of the soil. Then, according to the obtained regional train operation characteristics, the dynamic load of the train is set in the model to simulate the vibration, impact and other effects of the train on the soil during travel. By reasonably setting the model parameters and boundary conditions, such as train speed changes, load distribution, etc., the model can accurately reflect the dynamic interaction relationship between the train and the soil, thereby constructing the soil-train dynamic interaction model.

[0054] Based on the implicit time integration algorithm of finite element, the soil-pipe curtain interaction model and the soil-train dynamic interaction model are solved in the time domain respectively. For the soil-pipe curtain interaction model, according to the displacement boundary conditions and material constitutive relationship during the pipe curtain jacking process, the displacement of each node at different times under the action of the pipe curtain construction alone is calculated by iteratively solving the finite element equation, and then the surface settlement amount and settlement rate caused by the pipe curtain construction are obtained. For the soil-train dynamic interaction model, considering the change of the train moving load over time, the train load is equivalent to the node force applied to the soil model, and the implicit time integration algorithm is also used to solve the finite element equation, and the displacement of each node of the soil caused by the train operation is obtained. The change of the surface settlement amount and settlement rate under the action of the train operation alone is obtained. Then, the superposition principle is used to superimpose the settlement amount and settlement rate of the same node at the same time caused by the train operation and the pipe curtain construction. The difference between the surface settlement under the combined effect of train operation and pipe curtain construction and the surface settlement under pipe curtain construction alone is calculated to obtain the settlement increment; the difference between the settlement rates of the two is calculated to obtain the settlement rate increment, which is used as the dynamic compensation increment. Through this algorithm, the influence of the two factors of pipe curtain construction and train operation on the surface settlement is comprehensively considered, providing accurate data support for the subsequent adjustment of the static surface settlement prediction results.

[0055] In a possible implementation, step S330 further includes: Step S331: Compensating the static surface settlement prediction result according to the settlement increment and the settlement rate increment, and outputting the surface settlement prediction result.

[0056] Specifically, the obtained settlement increment and settlement rate increment are combined with the static surface settlement prediction results to obtain more accurate surface settlement prediction results. First, the settlement increment is added to the settlement data in the static surface settlement prediction results, which directly changes the predicted final surface settlement value, reflecting the impact of dynamic factors such as train operation on the total surface settlement. At the same time, according to the settlement rate increment, the static predicted settlement rate is adjusted. If the settlement rate increment is positive, the speed of settlement change over time is accelerated; if it is negative, the speed is slowed down. After such dual adjustment of settlement and settlement rate, the comprehensive impact of various factors such as pipe curtain construction and regional train operation on surface settlement is fully considered. Finally, the adjusted settlement, settlement rate and corresponding settlement distribution data are integrated and output to form a complete surface settlement prediction result, which provides an accurate basis for engineering personnel to evaluate the impact of pipe curtain construction on the surface, formulate construction plans and take corresponding measures.

[0057] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0058] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0059] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis, characterized in that: include: Combined with the regional soil property information, steel pipe property information and expected jacking construction parameters, simulation predictions of pipe-roof jacking impact, pipe-roof structure deflection and deformation, and soil arching effect are carried out to generate predicted jacking impact data, predicted deflection and deformation data, and predicted soil arching effect data; Based on the principle of integrated learning, the surface settlement of pipe-roof construction is predicted according to the predicted jacking influence data, the predicted deflection deformation data and the predicted soil arch effect data, and a static surface settlement prediction result is generated; Combined with regional train operation characteristics and regional soil property information, dynamic compensation analysis of surface settlement is performed to determine the dynamic compensation increment, adjust the static surface settlement prediction result, and output the surface settlement prediction result.

2. The method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis according to claim 1 is characterized in that: Obtain regional soil property information, steel pipe property information and expected jacking construction parameters, including: Collecting soil type, soil layer structure, soil mechanical properties and groundwater level in the area to be constructed as soil attribute information of the area; Collecting the material, strength, diameter and wall thickness of the steel pipe used for pipe curtain construction as the steel pipe attribute information; The expected jacking speed and expected jacking sequence of the steel pipe jacking operation are collected as the expected jacking construction parameters.

3. The method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis according to claim 2 is characterized in that: Combined with the regional soil property information, steel pipe property information and expected jacking construction parameters, the simulation prediction of the pipe curtain jacking impact and pipe curtain structure deflection deformation is carried out, including: Based on finite element analysis, a soil-pipe curtain interaction model is established according to the soil property information and the steel pipe property information of the region; The soil-pipe curtain interaction model is used to perform pipe curtain jacking simulation according to the expected jacking construction parameters, and predicted jacking impact data and predicted deflection deformation data are output, wherein the predicted jacking impact data includes first surface settlement prediction information, and the first surface settlement prediction information includes a first predicted settlement amount, a first predicted settlement rate, and a first predicted settlement distribution.

4. The method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis according to claim 3 is characterized in that: The soil-pipe-roof interaction model is established, including: Perform geometric feature modeling based on the soil attribute information and steel pipe attribute information of the region to generate a soil geometry model and a pipe curtain geometry model; Meshing the body geometric model and the pipe-roof geometric model according to a predetermined size to obtain a soil body mesh model and a pipe-roof mesh model; Based on finite element analysis, a soil-pipe-roof interaction model is established according to the soil grid model and the pipe-roof grid model.

5. The method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis according to claim 3 is characterized in that: The soil-pipe-roof interaction model is used to perform pipe-roof jacking simulation according to the expected jacking construction parameters, soil deformation prediction is performed, and predicted soil arching effect data is output, wherein the predicted soil arching effect data includes predicted arch pressure distribution.

6. The method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis according to claim 1 is characterized in that: Based on the principle of integrated learning, the surface settlement prediction of pipe-roof construction is carried out according to the predicted jacking influence data, the predicted deflection deformation data and the predicted soil arch effect data, and the static surface settlement prediction results are generated, including: Constructing Q surface subsidence prediction branches according to Q integrated learning operators, wherein the integrated learning operator includes at least a feedforward neural network and a random decision forest, and Q is an integer greater than or equal to 2; According to the historical pipe curtain construction records, sample data are collected to perform supervised training and verification training on the Q surface settlement prediction branches, and a surface settlement prediction model is integrated to be constructed; The predicted jacking influence data, predicted deflection deformation data and predicted soil arching effect data are input into the surface settlement prediction model, and a static surface settlement prediction result is output.

7. The method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis according to claim 6 is characterized in that: Collect sample data to perform supervised training and verification training on the Q surface subsidence prediction branches respectively, and integrate and construct a surface subsidence prediction model, including: According to the historical pipe curtain construction records, the sample jacking influence data set, the sample steel pipe deflection deformation data set and the sample soil arching effect data set were collected, and the sample surface settlement data under different sample jacking influence data, sample steel pipe deflection deformation data and sample soil arching effect data were obtained to construct the sample surface settlement data set, where the surface settlement data includes settlement amount, settlement rate and settlement distribution; The sample jacking influence data set, the sample steel pipe deflection deformation data set, the sample soil arch effect data set and the sample surface settlement data set are used as sample data and divided equally into Q parts, and then divided in proportion to obtain Q training sets and Q validation sets; Using the Q training sets and the Q validation sets, respectively, the Q surface subsidence prediction branches are supervised and validated until expected conditions are met, thereby obtaining Q converged surface subsidence prediction branches; Based on the weighted integration principle, the Q convergent surface subsidence prediction branches are combined to obtain the surface subsidence prediction model.

8. The method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis according to claim 4 is characterized in that: Combined with regional train operation characteristics and regional soil property information, dynamic compensation analysis of surface settlement is performed to determine the dynamic compensation increment, including: Acquire regional train operation characteristics, wherein the regional train operation characteristics include historical passing speed, historical load average, and expected passing times; Based on finite element analysis, a soil-train dynamic interaction model is constructed according to the regional train operation characteristics and regional soil property information; The soil-pipe-roof interaction model and the soil-train dynamic interaction model are used to perform dynamic compensation analysis of surface settlement to obtain dynamic compensation increments, wherein the dynamic compensation increments include settlement increments and settlement rate increments.

9. The method for predicting ground settlement during pipe-roof construction combined with multi-factor impact analysis according to claim 8, characterized in that: The static surface settlement prediction result is compensated according to the settlement increment and the settlement rate increment, and the surface settlement prediction result is output.

Images