A method for evaluating the structural state of a medium-low speed maglev transition section

Abstract

The application discloses a kind of low-speed maglev pile plate structure formula transition section state evaluation method, belong to low-speed maglev traffic engineering structural health monitoring technical field, the evaluation method includes: including following process: first, build vehicle-track-pile-road coupling model, simulation calculation is carried out under preset full working condition, the dynamic response data of pile plate structure is extracted and the health state under each working condition is marked, and full working condition sample database is constructed;Then based on the sample database training machine learning model, obtain the intelligent evaluation model of pile plate structure;Finally, the real-time dynamic response data of pile plate structure is collected, input intelligent evaluation model carries out real-time state diagnosis, and outputs health state diagnosis report.This method realizes the accurate, real-time intelligent evaluation of pile plate structure formula transition section state, provides reliable technical support for pile plate structure formula transition section operation and maintenance, improves operation and maintenance efficiency and structural safety, guarantees low-speed maglev train operation safety, with strong engineering practicability.

Description

Technical Field

[0001] This invention belongs to the field of structural health monitoring technology for medium- and low-speed maglev systems, specifically relating to a method for assessing the state of a transition section of a medium- and low-speed maglev pile-slab structure. Background Technology

[0002] With the development and advancement of electromagnetic control theory, maglev transportation has become a new mode of rail transportation following wheel-rail transportation. In medium- and low-speed maglev transportation systems, the transition section, as a key component connecting bridges and roadbeds of different foundation types, directly affects driving safety, smoothness, and ride comfort. Currently, the smoothness and safety testing of transition section structures mainly relies on periodic manual inspections, which are inefficient and highly subjective. Finite element analysis calculations are costly and difficult to perform in real-time assessment. Summary of the Invention

[0003] In response to one or more of the above-mentioned defects or improvement needs of the existing technology, the present invention provides a method for evaluating the state of a transition section of a medium- and low-speed maglev piling-slab structure, which can realize real-time intelligent evaluation of the structural state of the transition section.

[0004] To achieve the above objectives, this invention provides a method for evaluating the condition of a medium- and low-speed maglev piling-slab structure transition section. The piling-slab structure transition section, from top to bottom, comprises a track panel, a rail-bearing beam, a piling-slab structure, and a roadbed. The piling-slab structure includes a bearing plate and pile foundations. The evaluation method includes the following steps: (1) Build a vehicle-track-pile-road coupled model, perform simulation calculations on the model under preset full working conditions, extract dynamic response data of pile-slab structure, mark the health status level of pile-slab structure under each working condition, and construct a full working condition sample database of pile-slab structure. (2) Based on the full-condition sample database, train a machine learning model to obtain an intelligent evaluation model for pile-slab structures; (3) Collect real-time dynamic response data of the pile-slab structure, input it into the intelligent evaluation model, perform real-time status diagnosis of the pile-slab structure, and obtain a health status diagnosis report of the pile-slab structure.

[0005] As a further improvement of the present invention, process (1) includes the following steps: S11. Establish the finite element model of the pile-slab transition section; S12. Establish a model of the maglev train and track structure in multibody dynamics software, and calculate the time history of the dynamic force transmitted to the track structure when the maglev train runs at different speeds. S13. By sampling or full factorial experimental design, change the structural working condition parameters in S11 and the train operating condition parameters in S12, apply the dynamic force time history under the corresponding working condition to the finite element model, and complete the vehicle-track-pile-road coupled dynamic simulation of the healthy state under the full working condition corresponding to the experimental design. S14. Simulate the damage mode of the pile-slab structure in the finite element model, combine the damage parameters with various working condition parameters, and apply the dynamic force time history under the corresponding working condition to complete the vehicle-rail-pile-road coupled dynamic simulation of the damage state under the full working condition corresponding to the experimental design. S15. Based on the simulation results in steps S13 and S14, extract the original dynamic response data of the pile-slab structure under healthy and damaged states, extract high-dimensional feature vectors from the original dynamic response data, and label the corresponding health state levels to construct a full-condition sample database of the pile-slab structure.

[0006] As a further improvement of the present invention, the damage modes in step S14 include the development of microcracks at the bottom of the plate, the separation of the rail beam from the plate, uneven settlement of the pile foundation, and loss of pile integrity; the methods for simulating the damage modes of the pile-plate structure in the finite element model include reducing the elastic modulus of the material, introducing contact pair separation, and applying a forced displacement field.

[0007] As a further improvement of the present invention, the original dynamic response data in step S15 includes the original time history data of the acceleration response, dynamic strain response and dynamic displacement response of the pile-slab structure; the full-condition sample database includes the condition parameters, high-dimensional feature vectors and corresponding health status levels of each simulation.

[0008] As a further improvement of the present invention, the high-dimensional feature vector includes time-domain features, frequency-domain features, and time-frequency domain joint features extracted from the original time history data.

[0009] As a further improvement of the present invention, process (2) includes the following steps: S21. Preprocess the full-condition sample database and divide the preprocessed database into training set, validation set and test set. S22. Using the training set, train multiple different types of machine learning models in parallel. S23. Evaluate each trained machine learning model using the validation set, perform hyperparameter tuning on each machine learning model based on the validation set evaluation results, and construct a Stacking ensemble model using all tuned models as base learners. S24. Use the test set to verify the performance of all optimized machine learning models and Stacking ensemble models; and use SHAP analysis to quantify the contribution of each feature index to the model evaluation results, and verify the physical rationality of the decisions of all optimized machine learning models and Stacking ensemble models. S25. The final model with the best performance and verified by physical rationality shall be used as the intelligent evaluation model.

[0010] As a further improvement of the present invention, the various types of machine learning models include: a baseline model for processing tabular data, a deep model for processing long-sequence time-course data, and a robust model for small-sample scenarios.

[0011] As a further improvement of the present invention, the real-time dynamic response data is collected through an Internet of Things (IoT) sensor network, which includes an accelerometer, a fiber optic strain sensor, and a hydrostatic level deployed on the pile-slab structure.

[0012] As a further improvement of the present invention, the health status report includes a health status level and predicted values ​​of key performance indicators; when the health status level and / or the predicted values ​​of key performance indicators reach the warning value, the intelligent assessment model automatically sends alarm information and a preliminary diagnostic report to the operation and maintenance personnel.

[0013] As a further improvement to the present invention, this evaluation method also includes: (4) The adaptive evolution of the intelligent evaluation model is achieved based on incremental learning, which specifically includes the following steps: S41. Store the measured data verified on-site and the actual health status level corresponding to the pile-slab structure as incremental samples into the incremental database. S42. Periodically merge the incremental database with the full-condition sample database to retrain and verify the intelligent evaluation model, thereby updating and optimizing the intelligent evaluation model.

[0014] The aforementioned improved technical features can be combined with each other as long as they do not conflict with each other.

[0015] In summary, the beneficial effects of the above-described technical solutions conceived by this invention compared with the prior art include: (1) The method for assessing the state of the transition section of a medium- and low-speed maglev piling-slab structure according to the present invention includes the following process: First, a vehicle-track-pile-road coupled model is built, and simulation calculations are performed under preset full working conditions. The dynamic response data of the piling-slab structure is extracted and the health status under each working condition is marked to construct a full working condition sample database. Then, a machine learning model is trained based on the sample database to obtain an intelligent assessment model for the piling-slab structure. Finally, real-time dynamic response data of the piling-slab structure is collected, input into the intelligent assessment model for real-time state diagnosis, and a health status diagnosis report is output. This method realizes accurate and real-time intelligent assessment of the state of the piling-slab structure transition section, provides reliable technical support for the operation and maintenance of the piling-slab structure transition section, improves operation and maintenance efficiency and structural safety, ensures the safe operation of medium- and low-speed maglev trains, and has strong engineering practicality.

[0016] (2) The medium- and low-speed maglev pile-slab structure transition section state assessment method of the present invention establishes a finite element model of the pile-slab structure transition section, calculates the force transmitted from the train to the track structure during operation in multibody dynamics software, applies the force to the finite element model to realize vehicle-track-pile-road coupling, completes the vehicle-track-pile-road coupling dynamic simulation of the healthy state and damaged state under the full working condition corresponding to the experimental design, extracts the original data to build a sample library, realizes full working condition coverage of the pile-slab structure transition section, and improves the training efficiency and recognition accuracy of the subsequent intelligent assessment model.

[0017] (2) The medium- and low-speed maglev pile-slab structure transition section state assessment method of the present invention improves the damage identification accuracy and assessment reliability of the model by training multiple different types of machine learning models in parallel, combined with validation set hyperparameter tuning and Stacking integrated modeling strategy; at the same time, it quantifies the feature contribution in the model performance verification stage through SHAP analysis to verify the physical rationality of the model decision and avoid the defects of the "black box" model.

[0018] (4) The medium- and low-speed maglev pile-slab structure transition section state assessment method of the present invention is based on incremental learning to achieve adaptive updating of the model, which can adapt to the influence of factors such as long-term aging of structural materials and environmental changes, ensure the long-term effectiveness and stability of the assessment system, and reduce operation and maintenance costs. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of the method for evaluating the state of the transition section of the medium- and low-speed maglev pile-slab structure in an embodiment of the present invention; Figure 2 This is a cross-sectional view of the transition section of the pile-slab structure in an embodiment of the present invention; Figure 3 This is a dynamic simulation model of the pile-slab structure transition section in this embodiment of the invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0022] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0023] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0024] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0025] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can mean that the first feature is in direct contact with the second feature, or that the first feature is in indirect contact with the second feature through an intermediate medium. Furthermore, "above," "over," and "on top" of the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply that the first feature is at a lower horizontal level than the second feature.

[0026] Example: The preferred embodiment of the present invention provides a method for evaluating the state of a medium- and low-speed maglev pavement-slab transition section. The pavement-slab transition section being evaluated comprises, from top to bottom, a track panel, a rail-bearing beam, a pavement-slab structure, and a roadbed. The pavement-slab structure includes a bearing plate and pile foundations. Please refer to [link to relevant documentation]. Figure 1 The evaluation method includes the following process: (1) Build a vehicle-track-pile-road coupled model, perform simulation calculations on the model under preset full working conditions, extract dynamic response data of pile-slab structure, mark the health status level of pile-slab structure under each working condition, and construct a full working condition sample database of pile-slab structure. (2) Based on the full-condition sample database, train a machine learning model to obtain an intelligent evaluation model for pile-slab structures; (3) Collect real-time dynamic response data of the pile-slab structure, input it into the intelligent evaluation model, perform real-time status diagnosis of the pile-slab structure, and obtain a health status diagnosis report of the pile-slab structure.

[0027] Preferably, process (1) includes the following steps: S11. Establish the finite element model of the pile-slab transition section; S12. Establish a model of the maglev train and track structure in multibody dynamics software, and calculate the time history of the dynamic force transmitted to the track structure when the maglev train runs at different speeds. S13. By sampling or full factorial experimental design, change the structural working condition parameters in S11 and the train operating condition parameters in S12, apply the dynamic force time history under the corresponding working condition to the finite element model, and complete the vehicle-track-pile-road coupled dynamic simulation of the healthy state under the full working condition corresponding to the experimental design. S14. Simulate the damage mode of the pile-slab structure in the finite element model, combine the damage parameters with various working condition parameters, and apply the dynamic force time history under the corresponding working condition to complete the vehicle-rail-pile-road coupled dynamic simulation of the damage state under the full working condition corresponding to the experimental design. S15. Based on the simulation results in steps S13 and S14, extract the dynamic response data of the pile-slab structure under healthy and damaged states, extract high-dimensional feature vectors from the dynamic response data, label the corresponding health state levels, and construct a full-condition sample database of the pile-slab structure.

[0028] Preferably, the finite element software used to establish the finite element model in step S11 is any one of ANSYS, ABAQUS, or Midas.

[0029] In a specific preferred embodiment, such as Figure 2 As shown, the pile-slab transition section, from top to bottom, includes a track panel 1, a track-bearing beam 2, a pile-slab structure 3, and a roadbed 4. The track-bearing beam 2 is a reinforced concrete structure, with its upper part connected to the track panel 1 and its lower part inserted into the roadbed 4 to a certain depth. A finite element method was established using this method. Figure 2 The finite element model of the pile-slab transition section shown is based on C3D8R solid elements. Specific structural parameters are as follows: the rail support beam is made of C40 reinforced concrete with a density of 2500 kg / m³. 3 The elastic modulus is 3.25 × 10⁻⁶. 4 MPa, Poisson's ratio is 0.3. The track panel is made of steel with a density of 7800 kg / m³. 3 The elastic modulus is 2.059 × 10⁻⁶. 5 MPa, Poisson's ratio is 0.15. The subgrade structure from top to bottom consists of backfill soil, surface layer of subgrade, bottom layer of subgrade, and soil layer below the subgrade; the backfill soil is improved soil with a density of 1850 kg / m³. 3 The elastic modulus is 0.25 × 10⁻⁶. 3 MPa; the surface layer of the subgrade consists of graded crushed stone, 0.3 m thick, with a density of 2400 kg / m³. 3 The elastic modulus is 1.58 × 10⁻⁶. 3 MPa; the bottom layer of the substrate uses group A and B fillers, with a thickness of 0.4 m and a density of 2350 kg / m³. 3 The elastic modulus is 0.53 × 10⁻⁶. 3 MPa. The pile-slab structure consists of a bearing slab and piles, using C30 concrete with a density of 2400 kg / m³. 3 The elastic modulus is 3×10 4 MPa, where the bearing plate is 0.8 m thick and 11.6 m long; the pile diameter is 1 m and the pile length is 9 m. Boundary conditions employ a combination of fixed and elastic constraints to ensure consistency between the model and the actual structure, such as... Figure 3 As shown, one side of the pile-slab structure 3 is connected to the ordinary roadbed section, and the other side is connected to the bridge abutment.

[0030] Preferably, the multibody dynamics software in step S12 is any one of SIMPACK, ADAMS, or UM. Preferably, the dynamic force time history transmitted by the maglev train to the track structure includes the dynamic levitation force time history and the guiding force time history.

[0031] Specifically, preferably, it is established in multibody dynamics software. Figure 3 The rigid-flexible coupling between the maglev train 5 and the track structure is implemented by calculating the dynamic levitation force time history and guiding force time history transmitted from the suspension frame to the track structure when the train 5 passes through the track structure at various speeds. This force time history is then used as a dynamic load and applied to the track panel 1 of the finite element model established in step S11. The simulation time step is set, and a coupled dynamic simulation of the vehicle-track-pile-road is performed to simulate the vibration response of the pile-slab transition section during train operation. Preferably, the track structure includes the track panel.

[0032] More preferably, the structural operating parameters in step S11 include the geometric parameters, material parameters, boundary condition parameters, and contact parameters of the pile-slab transition section; the train operating parameters in step S12 include train speed, axle load, and acceleration / deceleration characteristics.

[0033] Preferably, the damage modes in step S14 include the development of microcracks at the bottom of the slab, separation of the track beam from the slab, uneven settlement of the pile foundation, and loss of pile integrity. The quantitative simulation of the damage modes includes reducing the elastic modulus of the material, introducing contact pair separation, and applying a forced displacement field. For example, microcracks are achieved by reducing the elastic modulus of the concrete at the bottom of the slab (e.g., 10%-50%), separation is achieved by introducing contact pair separation between the track beam and the slab, uneven settlement of the pile foundation is achieved by applying forced displacement (e.g., 2-10 mm) to the pile bottom, and loss of pile integrity is achieved by reducing the local elastic modulus of the pile body (e.g., 20%-60%). Further, for each damage mode, multiple damage degrees (e.g., mild, moderate, severe) and multiple spatial locations (e.g., mid-span, slab end, pile top) can be set, combined with multiple sets of normal working condition parameters in step S13, to obtain a large number of damage conditions. Transient dynamic analysis is performed on each set of damage conditions to obtain dynamic response data under the damage state.

[0034] Preferably, the original dynamic response data in step S15 includes the original time history data of the acceleration response, dynamic strain response, and dynamic displacement response of the pile-slab structure; the full-condition sample database contains condition parameters, high-dimensional feature vectors, and corresponding health status levels. Specifically, the original time history data of the acceleration response, dynamic strain response, and dynamic displacement response of the pile-slab structure are extracted from the health status model in S13 and the damage status model in S14; feature extraction is performed on the original time history data to obtain time-domain features, frequency-domain features, and time-frequency-domain joint features, and a high-dimensional feature vector is constructed; the condition parameters, high-dimensional feature vectors, and corresponding health status levels (e.g., healthy = 0, microcracks = 1, voids = 2, settlement = 3, pile damage = 4) are integrated to construct a structured full-condition sample database TSYDATA.

[0035] More preferably, process (2) includes the following steps: S21. Preprocess the full-condition sample database and divide the preprocessed database into training set, validation set and test set. S22. Using the training set, train multiple different types of machine learning models in parallel. S23. Evaluate each trained machine learning model using the validation set, perform hyperparameter tuning on each machine learning model based on the validation set evaluation results, and construct a Stacking ensemble model using all tuned models as base learners. S24. Use the test set to verify the performance of all optimized machine learning models and Stacking ensemble models; and use SHAP analysis to quantify the contribution of each feature index to the model evaluation results, and verify the physical rationality of the decisions of all optimized machine learning models and Stacking ensemble models. S25. The final model with the best performance and verified by physical rationality shall be used as the intelligent evaluation model.

[0036] Specifically, preferably, in step S21, the full-condition sample database is preprocessed and divided into a training set, a validation set, and a test set in a 7:2:1 ratio. In step S22, after standardizing the input features of the training set, multiple different types of machine learning models are trained in parallel. Preferably, the multiple different types of machine learning models include a baseline model for processing tabular data, a deep model for processing long-sequence time-history data, and a robust model for small-sample scenarios. In step S23, hyperparameter tuning employs grid search or Bayesian optimization. Preferably, the intelligent assessment model outputs a health status report along with the confidence level of the assessment results.

[0037] More preferably, the real-time dynamic response data is collected from the transition section structure via an Internet of Things (IoT) sensor network, which includes accelerometers, fiber optic strain sensors, and hydrostatic levels deployed on the pile-slab structure.

[0038] Preferably, the health status report includes a health status level and predicted values ​​of key performance indicators. When the health status level and / or the predicted value of the key performance indicators reach the warning value, the intelligent assessment model automatically sends alarm information and a preliminary diagnostic report to the operation and maintenance personnel.

[0039] For example, health status levels include L1-Excellent, L2-Good, L3-Concern, L4-Abnormal, and L5-Hazardous. Key performance indicators include settlement, deformation, etc. Preferably, the intelligent assessment model automatically triggers four levels of early warning (green, yellow, orange, and red) based on preset thresholds of the health status level and / or predicted values ​​of key performance indicators. When an orange or higher warning is triggered, an alarm message and a preliminary diagnostic report are sent to the maintenance personnel to remind them to take timely action.

[0040] Preferably, the method for assessing the state of the transition section of a medium- and low-speed maglev piling-slab structure further includes the following process: (4) The adaptive evolution of the intelligent evaluation model is achieved based on incremental learning, which specifically includes the following steps: S41. The measured data verified on-site and the actual health status level corresponding to the pile-slab structure are used as incremental samples and stored in the incremental database. S42. Periodically merge the incremental database with the full-condition sample database to retrain and verify the intelligent evaluation model, thereby updating and optimizing the intelligent evaluation model.

[0041] In this embodiment, the evaluation system is self-evolved through real incremental samples, enabling it to continuously adapt to the long-term aging of structural materials and environmental changes.

[0042] The method for assessing the condition of the transition section of a medium- and low-speed maglev piling-slab structure in this invention achieves full coverage of the transition section structure under all working conditions, accurate damage identification, and real-time condition diagnosis. It solves the problems of low efficiency, insufficient accuracy, and poor real-time performance of existing assessment methods, ensuring the safe operation of medium- and low-speed maglev trains and has strong engineering applicability.

[0043] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for assessing the condition of a medium-low speed maglev pavement-slab transition section, wherein the pavement-slab transition section comprises, from top to bottom, a track panel, a track-bearing beam, a pavement-slab structure, and a roadbed, wherein the pavement-slab structure comprises a bearing plate and pile foundations; characterized in that, The process includes the following: (1) Build a vehicle-track-pile-road coupled model, perform simulation calculations on the model under preset full working conditions, extract dynamic response data of pile-slab structure, mark the health status level of pile-slab structure under each working condition, and construct a full working condition sample database of pile-slab structure. (2) Based on the full-condition sample database, train a machine learning model to obtain an intelligent evaluation model for pile-slab structures; (3) Collect real-time dynamic response data of the pile-slab structure, input it into the intelligent evaluation model, perform real-time status diagnosis of the pile-slab structure, and obtain a health status diagnosis report of the pile-slab structure.

2. The method for assessing the state of the transition section of a medium- and low-speed maglev pile-slab structure according to claim 1, characterized in that, Process (1) includes the following steps: S11. Establish the finite element model of the pile-slab transition section; S12. Establish a model of the maglev train and track structure in multibody dynamics software, and calculate the time history of the dynamic force transmitted to the track structure when the maglev train runs at different speeds. S13. By sampling or full factorial experimental design, change the structural working condition parameters in S11 and the train operating condition parameters in S12, apply the dynamic force time history under the corresponding working condition to the finite element model, and complete the vehicle-track-pile-road coupled dynamic simulation of the healthy state under the full working condition corresponding to the experimental design. S14. Simulate the damage mode of the pile-slab structure in the finite element model, combine the damage parameters with various working condition parameters, and apply the dynamic force time history under the corresponding working condition to complete the vehicle-rail-pile-road coupled dynamic simulation of the damage state under the full working condition corresponding to the experimental design. S15. Based on the simulation results in steps S13 and S14, extract the original dynamic response data of the pile-slab structure under healthy and damaged states, extract high-dimensional feature vectors from the original dynamic response data, and label the corresponding health state levels to construct a full-condition sample database of the pile-slab structure.

3. The method for assessing the state of the transition section of a medium- and low-speed maglev pile-slab structure according to claim 2, characterized in that, The damage modes in step S14 include the development of microcracks at the bottom of the plate, separation of the rail beam from the plate, uneven settlement of the pile foundation, and loss of pile integrity. The methods for simulating the damage modes of the pile-plate structure in the finite element model include reducing the elastic modulus of the material, introducing contact pair separation, and applying a forced displacement field.

4. The method for assessing the state of the transition section of a medium- and low-speed maglev pile-slab structure according to claim 2, characterized in that, The raw dynamic response data in step S15 includes the raw time history data of the acceleration response, dynamic strain response and dynamic displacement response of the pile-slab structure; the full-condition sample database includes the condition parameters, high-dimensional feature vectors and corresponding health status levels of each simulation.

5. The method for evaluating the state of the transition section of a medium- and low-speed maglev pile-slab structure according to claim 4, characterized in that, The high-dimensional feature vector includes time-domain features, frequency-domain features, and joint time-frequency-domain features extracted from the original time history data.

6. The method for assessing the state of the transition section of a medium-low speed maglev piling-slab structure according to any one of claims 1 to 5, characterized in that, Process (2) includes the following steps: S21. Preprocess the full-condition sample database and divide the preprocessed database into training set, validation set and test set. S22. Using the training set, train multiple different types of machine learning models in parallel. S23. Evaluate each trained machine learning model using the validation set, perform hyperparameter tuning on each machine learning model based on the validation set evaluation results, and construct a Stacking ensemble model using all tuned models as base learners. S24. Use the test set to verify the performance of all optimized machine learning models and Stacking ensemble models; and use SHAP analysis to quantify the contribution of each feature index to the model evaluation results, and verify the physical rationality of the decisions of all optimized machine learning models and Stacking ensemble models. S25. The final model with the best performance and verified by physical rationality shall be used as the intelligent evaluation model.

7. The method for assessing the state of the transition section of a medium- and low-speed maglev piling-slab structure according to claim 6, characterized in that, The various types of machine learning models include: a baseline model for processing tabular data, a deep model for processing long-sequence time-course data, and a robust model for small-sample scenarios.

8. The method for assessing the state of the transition section of a medium-low speed maglev piling-slab structure according to any one of claims 1 to 5 and 7, characterized in that, The real-time dynamic response data is collected through an Internet of Things (IoT) sensor network, which includes accelerometers, fiber optic strain sensors, and hydrostatic levels deployed on the pile-slab structure.

9. The method for assessing the state of the transition section of a medium-low speed maglev piling-slab structure according to any one of claims 1 to 5 and 7, characterized in that, The health status report includes the health status level and predicted values ​​of key performance indicators; when the health status level and / or the predicted value of the key performance indicators reach the warning value, the intelligent assessment model automatically sends alarm information and a preliminary diagnostic report to the operation and maintenance personnel.

10. The method for assessing the state of the transition section of a medium-low speed maglev piling-slab structure according to any one of claims 1 to 5 and 7, characterized in that, It also includes the following processes: (4) The adaptive evolution of the intelligent evaluation model is achieved based on incremental learning, which specifically includes the following steps: S41. The measured data verified on-site and the actual health status level corresponding to the pile-slab structure are used as incremental samples and stored in the incremental database. S42. Periodically merge the incremental database with the full-condition sample database to retrain and verify the intelligent evaluation model, thereby updating and optimizing the intelligent evaluation model.

Images