Tunnel lining multi-scale load simulation system and method based on flexible grid

Abstract

The invention discloses a tunnel lining multi-scale load simulation system and method based on a flexible grid, and relates to the technical field of tunnel engineering detection, and the system comprises the components of a flexible loading grid which is used for covering a load action road surface of a tunnel reduced scale model and comprises a plurality of modularized magnetic grid units; a micro electromagnetic vibration actuator array is integrated in each grid unit, and each micro electromagnetic vibration actuator in the array acts an electromagnetic pulse signal on an internal coil to drive and generate mechanical vibration; the pulse generator is used for generating electromagnetic pulse signals and sending the electromagnetic pulse signals to the miniature electromagnetic vibration actuators; and the intelligent control module is used for converting the set load value into an electromagnetic pulse parameter value and controlling the generation of a plurality of electromagnetic pulse signals. The designed grid units can be magnetically attracted and spliced according to road structures in different tunnels, electromagnetic pulse vibration can be loaded through dynamic partition, and the dynamic load change of the tunnel in a complex stress environment is accurately simulated.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tunnel engineering detection, and in particular to a tunnel lining multi-scale load simulation system and method based on a flexible grid. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute the prior art.

[0003] In tunnel engineering, the lining structure is subjected to the pressure of surrounding rock, groundwater and traffic load for a long time, which is prone to cause cracks, leakage and other diseases. Among them, the traffic load acting on the internal road of the lining is the core dynamic load in the operation stage, and the long-term repeated action of the traffic load on the lining structure is the key factor leading to structural cumulative damage. Accurate simulation of the application characteristics of the traffic load is crucial to the study of the mechanical behavior and performance evolution of the lining.

[0004] At present, in the simulation and application of traffic load of tunnel model structure, the existing simulation devices or systems have many shortcomings: the current disease loading model test device for subway tunnel structure uses a single exciter to simulate traffic load, which is a single-point loading method and cannot accurately simulate multi-point complex traffic load; moreover, the current model test device for cyclic dynamic load soil settlement and the test device for cumulative damage of tunnel bottom structure under cyclic load all load the load on the overlying soil of the lining, and then convert the mechanical energy into stress to pressurize the tunnel lining structure. This step involves multiple energy conversions, and energy loss is inevitable, which makes it difficult to accurately control the load applied to the tunnel lining structure, cannot stably and accurately simulate the change characteristics of dynamic traffic load, and greatly limits its application in diversified tunnel model research. SUMMARY

[0005] To solve the above problems of the prior art, the present application provides a tunnel lining multi-scale load simulation system and method based on a flexible grid, which is designed by a flexible loading grid composed of a plurality of modular magnetic grid units. The grid units can be magnetically spliced according to different internal road structures of the tunnel lining, and a micro electromagnetic vibration actuator is integrated in the grid unit. By dynamically partitioning and loading electromagnetic pulse vibration, the dynamic load change of the tunnel in a complex stress environment can be accurately simulated, the fatigue damage analysis of the tunnel lining can be realized, the problems of inaccurate load simulation and large energy loss of the existing simulation system can be solved, the effective load simulation can be realized, and the foundation for subsequent accurate damage analysis is laid.

[0006] In a first aspect, the present application provides a tunnel lining multi-scale load simulation system based on a flexible grid.

[0007] A tunnel lining multi-scale load simulation system based on a flexible grid, comprising: A flexible loading grid for covering the load-acting road surface of a tunnel scale model, comprising a plurality of modular magnetic grid units; each grid unit internally integrates an array of micro electromagnetic vibration actuators, and each micro electromagnetic vibration actuator in the array acts an electromagnetic pulse signal to an internal coil to drive mechanical vibration; A pulse generator for generating electromagnetic pulse signals and sending them to each micro electromagnetic vibration actuator; An intelligent control module for converting a set load value into an electromagnetic pulse parameter value and controlling the generation of multiple electromagnetic pulse signals.

[0008] The further technical solution further comprises a signal acquisition unit; The signal acquisition unit comprises a micro acceleration sensor and a pressure sensor embedded in the bottom of each grid unit, for monitoring vibration load data applied in the grid unit in real time; wherein the vibration load data includes the pulse peak value, the pulse repetition frequency of the vibration load, and the pulse pressure peak value acting on the internal road surface of the lining; The intelligent control module is further configured to adaptively adjust the electromagnetic pulse load distribution according to the set load value and the monitored vibration load data, and control the generation of multiple electromagnetic pulse signals.

[0009] The further technical solution, the electromagnetic pulse load distribution includes the pulse peak value, the repetition frequency, the pulse width, the pulse timing and the duration of the electromagnetic pulse signal applied to each grid unit.

[0010] The further technical solution, each modular magnetic grid unit is made of a silica gel-aramid composite material, and a permanent magnet array is embedded in the periphery of each grid unit for magnetically splicing multiple modular magnetic grid units to form a flexible loading grid.

[0011] The further technical solution, a silica gel damping vibration isolation ring is uniformly arranged around the periphery of each modular magnetic grid unit.

[0012] The further technical solution, the signal acquisition unit further comprises a fiber Bragg grating sensor arranged on the surface of the tunnel lining, for real-time acquisition of actual tunnel lining strain data; The intelligent control module is further configured to inversely deduce an optimal electromagnetic pulse load distribution based on a pre-loaded adaptive algorithm based on a digital twin model according to the acquired actual lining strain data, to regulate the generation of multiple electromagnetic pulse signals.

[0013] The further technical solution, according to the acquired actual lining strain data, the optimal electromagnetic pulse load distribution is inversely deduced by a pre-loaded adaptive algorithm based on a digital twin model, comprising: A digital twin model of a tunnel lining and its load action environment is constructed; The actual lining strain data collected in real time is input into the digital twin model of the tunnel lining, the target strain corresponding to the lining under different preset electromagnetic pulse load distributions in the digital twin model is compared with the actual strain, and whether to adjust the electromagnetic pulse load parameters is determined according to the comparison result; if the determination result is yes, the preloaded adaptive algorithm is used to make the actual strain close to the target strain as an optimization target, and the electromagnetic pulse load parameters required to be adjusted, i.e., the optimal electromagnetic pulse load distribution obtained by backstepping, are calculated.

[0014] In a second aspect, the present application provides a flexible mesh-based tunnel lining multi-scale load simulation method.

[0015] The flexible mesh-based tunnel lining multi-scale load simulation method is implemented based on the flexible mesh-based tunnel lining multi-scale load simulation system proposed in the first aspect, and includes the following steps: The model parameters are designed based on the similarity theory, the tunnel scale model is built, and the load value applied to the road surface inside the tunnel lining is preliminarily determined; The flexible loading mesh is covered on the load action road surface of the tunnel scale model, and the output load of each modular magnetic attraction mesh unit is calibrated; The preliminarily determined load value is converted into electromagnetic pulse parameter value, a plurality of electromagnetic pulse signals are generated and sent to the micro electromagnetic vibration actuators in each mesh unit to simulate the load; The actual tunnel lining strain data are collected in real time, the optimal electromagnetic pulse load distribution is backstepped by the adaptive algorithm based on the digital twin model, and the application of the simulated load is continuously adjusted.

[0016] In a further technical solution, the calibration of the output load of each modular magnetic attraction mesh unit is as follows: The vibration load data are monitored in real time; wherein the vibration load data include the pulse peak value, the pulse repetition frequency of the vibration load, and the pulse pressure peak value acting on the road surface inside the lining; The electromagnetic pulse load distribution is adaptively adjusted according to the set load value and the monitored vibration load data, and the generation of the plurality of electromagnetic pulse signals is controlled; wherein the electromagnetic pulse load distribution includes the pulse peak value, the repetition frequency, the pulse width, the pulse timing and the duration of the electromagnetic pulse signals applied to each mesh unit.

[0017] In a further technical solution, the load simulation is performed based on the tunnel scale model for a set number of times of load loading, the strain data after the application of the simulated load are collected, and the lining fatigue damage is calculated according to the collected data.

[0018] The above one or more technical solutions have the following beneficial effects: 1. The application provides a tunnel lining multi-scale load simulation system and method based on a flexible grid, which is used for a tunnel lining structure, and a flexible loading grid composed of a plurality of modular magnetic grid units is designed to cover a road surface subjected to a load, an electromagnetic pulse vibration driving mode is introduced into the grid, a micro electromagnetic vibration actuator is integrated into the grid unit, an electromagnetic pulse signal is converted into a mechanical vibration load by the electromagnetic vibration actuator, dynamic partition loading of the electromagnetic pulse vibration is realized, dynamic load changes of the tunnel in a complex stress environment can be accurately simulated, fatigue damage analysis of the tunnel lining can be carried out according to collected tunnel lining strain data under the dynamic load, the problems of inaccurate load simulation and large energy loss of the existing simulation system are solved, and the accuracy of load simulation and damage analysis is ensured. The high-frequency electromagnetic pulse vibration simulation function can accurately reproduce the instantaneous impact characteristics of the traffic load, the pulse current vibration driving mode is adopted, the energy conversion efficiency is relatively high, the vibration parameters (frequency, amplitude, etc.) are independently adjustable, and the simulation accuracy is higher.

[0019] 2. In the application, the modular magnetic grid units in the flexible loading grid can be magnetically spliced according to different road structures in tunnels, each grid unit is independently controlled, a multi-lane differentiated pulse load scene can be constructed, dynamic partition loading of the vibration load is realized, and accurate simulation of the multi-scale load of the tunnel lining is completed.

[0020] 3. In the application, test data is stored based on the block chain technology, the data is ensured to be tamper-proof, and a reliable digital asset library can be provided for intelligent operation and maintenance of the tunnel.

[0021] The advantages of the additional aspects of the application will be partially given in the following description, partially become obvious from the following description, or be known by the practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0022] The drawings accompanying the specification of the application form part of the application and serve to provide further understanding of the application, the illustrative embodiments of the application and the description thereof serve to explain the application and do not constitute an improper limitation of the application.

[0023] Figure 1 A schematic view of a 30x40cm flexible loading grid formed by magnetic splicing in the embodiment of the application; Figure 2 A structural schematic view of a single modular magnetic grid unit in the embodiment of the application; Figure 3 A schematic view of a micro electromagnetic pulse vibration actuator in the embodiment of the application; Figure 4 A schematic view of a tunnel lining multi-scale load simulation system based on a flexible grid in the embodiment of the application.

[0024] Wherein, 1, permanent magnet array; 2, vibration isolation ring; 3, micro electromagnetic vibration actuator array; 4, micro electromagnetic vibration actuator; 5, threading hole; 6, coil; 7, spring; 8, first resonator; 9, second resonator; 10, acceleration sensor; 11, pressure sensor; 12, first data line; 13, second data line; 14, third data line; 15, pulse generator; 16, intelligent control module. DETAILED DESCRIPTION

[0025] It should be noted that the following detailed description is exemplary only and is intended to provide further description of the present application in connection with the exemplary embodiments according to the present application, and is not intended to limit the exemplary embodiments according to the present application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. Furthermore, it should be understood that the use of the terms "including", "comprising", "having" and / or "containing" herein, are meant to be inclusive.

[0026] Example One The present embodiment proposes a tunnel lining multi-scale load simulation system based on a flexible grid, which can simulate the traffic load on the road inside the tunnel lining structure, and through the collection of the corresponding simulation information, can effectively analyze the stress and strain, vibration and the like of each part of the tunnel lining structure, thereby laying a foundation for the study of the performance evolution of the lining structure and the like.

[0027] Specifically, the simulation system proposed in the present embodiment includes a flexible loading grid, a pulse generator and an intelligent control module. The flexible loading grid, as an important component for loading, is formed by magnetically splicing a plurality of modular magnetic grid units, as shown in Figure 1 The modular magnetic grid units can be magnetically spliced according to different road structures inside the tunnel, and the flexible loading grid formed by the splicing is covered on the load-acting road surface inside the lining structure of the tunnel scale model (i.e. the load-acting road surface), and the inside of each grid unit is integrated with a 4x4 micro electromagnetic vibration actuator array 3, as shown in Figure 3 Each micro electromagnetic vibration actuator 4 in the array acts an electromagnetic pulse signal to the internal coil 6 to drive mechanical vibration, and each modular magnetic grid unit can be controlled individually, so as to realize distributed loading and simulation of the corresponding load, such as the dynamic load of 0-1kN.

[0028] As shown in Figure 2As shown, each modular magnetic grid unit is made of silica gel-aramid composite material, which is a flexible grid unit, and a Halbach permanent magnet array 1 is embedded in the four corners of each grid unit (100x100mm), and the single-side magnetic attraction force of the permanent magnet array is ≥200N, supporting curved surface adaptive splicing, and a flexible loading grid is formed by splicing multiple modular magnetic grid units through magnetic attraction. Preferably, silica gel damping vibration isolation rings 2 are uniformly arranged on the periphery of each modular magnetic grid unit and the magnetic attraction connection thereof, so as to reduce the mutual interference introduced by the superposition of adjacent unit vibration waves. In addition, a threading hole 5 is also provided on each modular magnetic grid unit to facilitate the threading and threading of the wire.

[0029] As shown, Figure 3 As shown, a plurality of resonators are embedded in each grid unit, and in this embodiment, a first resonator 8 and a second resonator 9 are provided, which are connected to the coil 6 of the micro electromagnetic vibration actuator. Preferably, the resonator is also connected to the coil 6 through a plurality of springs to serve as elastic support.

[0030] As shown, Figure 4 As shown, in the flexible loading grid, the micro electromagnetic vibration actuator integrated in each grid unit is connected to an external pulse generator 15 through the first data line 12 passing through the threading hole 5, and the pulse generator 15 is used to generate electromagnetic pulse signals and send them to each micro electromagnetic vibration actuator, which supports adjustable waveform (half-sine wave, rectangular wave, triangular wave, etc.), pulse width setting (50-500μs), and peak current control (0-5A). In addition, the pulse generator 15 is electrically connected to the intelligent control module 16, and the two communicate with each other, and the intelligent control module 16 is used to convert the set load value into electromagnetic pulse parameter value and control the generation of multiple electromagnetic pulse signals, so as to realize distributed loading and simulation of the load.

[0031] Specifically, the intelligent control module 16 controls the pulse generator 15 to generate electromagnetic pulse signals, which are transmitted to the coil 6 in the corresponding micro electromagnetic vibration actuator through the first data line 12 and the threading hole 5, and a changing current is generated in the coil. According to the law of electromagnetic induction, the changing current will generate a changing magnetic field around the coil, which interacts with the magnetic field generated by the permanent magnet array 1, thereby generating a periodic electromagnetic force on the coil; after the coil is subjected to the electromagnetic force, the resonator connected thereto is driven to move. Among them, the spring 7 not only has the function of elastic support, but also has the function of auxiliary resonance, which can make the resonator vibrate at a certain frequency and amplitude under the drive of the electromagnetic force, so as to convert the electromagnetic pulse signal into mechanical vibration and transmit it to the road surface inside the tunnel lining, thereby realizing the application of the load.

[0032] As an embodiment, as shown, Figure 3As shown, the system further comprises a signal acquisition unit, which comprises a micro acceleration sensor 10 and a pressure sensor 11 embedded in the bottom of each grid unit, which is used to monitor the vibration load data applied in the grid unit in real time, i.e. the pulse peak value, pulse repetition frequency of the vibration load and the pulse pressure peak value acting on the road surface inside the lining.

[0033] The acceleration sensor 10 and the pressure sensor 11 are electrically connected to the intelligent control module 16, i.e. the acceleration sensor 10 is connected to the external intelligent control module 16 through the second data line 13, and the pressure sensor is connected to the external intelligent control module 16 through the third data line 14. The intelligent control module 16 can dynamically adjust the electromagnetic pulse load distribution according to the set load value and the received monitoring vibration load data (such as pressure, etc.), i.e. adjust the pulse peak value, repetition frequency, pulse width, pulse timing and duration, etc. of the electromagnetic pulse signal applied to each grid unit, control the generation of multiple electromagnetic pulse signals, through this feedback adjustment mechanism, the pressure released by the micro electromagnetic vibration actuator 4 meets the user's demand, and the accuracy of load application and simulation is guaranteed.

[0034] As another embodiment, the above-mentioned signal acquisition unit further comprises a fiber grating sensor arranged on the surface of the tunnel lining, such as the left and right arch feet, left and right arch waist, arch top, etc., which can collect lining strain data (such as strain force, etc.) in real time and transmit these collected data to the intelligent control module. The intelligent control module can reversely deduce the optimal electromagnetic pulse load distribution according to the collected lining strain data through the self-adaptive algorithm based on the digital twin model pre-installed in it, and the process is as follows: First, a digital twin model of the tunnel lining and its load action environment is constructed, which can accurately map the geometric shape, material properties, mechanical behavior, etc. of the actual tunnel lining.

[0035] Then, the intelligent control module inputs the collected actual lining strain data into the digital twin model of the tunnel lining, and pre-sets the corresponding target strain of the lining under different electromagnetic pulse load distributions in the digital twin model, i.e. the reasonable strain of the lining under the action of real traffic load, etc. At this time, the module compares the actual collected strain data with the target strain data in the digital twin model, and adjusts the electromagnetic pulse load parameters according to the comparison result.

[0036] Specifically, when there is a deviation between the actual strain and the target strain, a pre-loaded adaptive algorithm is used, such as a model reference adaptive control (MRAC) algorithm, an adaptive filtering algorithm, an adaptive algorithm based on machine learning (such as reinforcement learning, neural networks, etc.), and the like, with the optimization goal of "making the actual strain as close as possible to the target strain", to calculate the required adjustment of the electromagnetic pulse load parameters (such as the pulse peak value, repetition frequency, pulse width, pulse timing and duration of the electromagnetic pulse in each modular magnetic attraction grid unit, etc.).

[0037] Finally, the intelligent control module generates and issues the adjusted pulse signal based on the above parameters, that is, according to the calculated adjustment amount, new electromagnetic pulse load parameters are generated, and these parameters are sent to the pulse generator, and the pulse generator generates an adjusted electromagnetic pulse signal to control the micro electromagnetic vibration actuator, thereby achieving adjustment of the electromagnetic pulse load distribution.

[0038] Through the above scheme, the difference between load simulation and actual situation can be effectively made up, and complex dynamic load can be accurately simulated to achieve precise damage assessment. That is: The load is set before the start of load simulation, but due to the production of the tunnel scale model, material properties (such as the actual elastic modulus and Poisson's ratio of concrete), boundary conditions (such as the simulation of surrounding rock), etc., there may be some deviation from the theoretical setting. By backstepping the optimal load distribution, these deviations can be effectively made up, making the load simulation more realistic; The actual tunnel bears complex and variable traffic loads, such as vehicle speed, load, distribution, etc., which will change dynamically. It is difficult to accurately simulate this dynamic complexity by relying solely on pre-set loads, while backstepping the optimal load can dynamically adjust the load distribution based on real-time strain data, more realistically restoring the dynamic characteristics of the actual load; The fatigue damage of the lining is closely related to the action of the load. Only when the load simulation is accurate enough, the fatigue damage calculation or prediction based on it (such as predicting the fatigue damage of the lining through the Miner fatigue accumulation rule or through a pre-trained deep neural network model reflecting the relationship between electromagnetic pulse load-lining strain-crack propagation) will be accurate. Backstepping the optimal load distribution can make the load simulation more accurate, thereby improving the reliability of damage assessment and providing a more scientific basis for the maintenance and reinforcement of tunnel lining; In addition, in long-term load simulation tests, the mechanical properties of the tunnel lining may change, such as the appearance of small damage leading to changes in stiffness, while backstepping the optimal load distribution can adaptively adjust the loading parameters to ensure the accuracy and effectiveness of the load simulation throughout the test process, rather than relying solely on the initial set load unchanged.

[0039] Example Two The embodiment provides a tunnel lining multi-scale load simulation method based on a flexible grid, and is realized based on a tunnel lining multi-scale load simulation system based on a flexible grid provided in embodiment one, and specifically includes the following steps. Step S1, model parameters are designed based on similarity theory, a tunnel scale model is built, and a load value applied to a road surface inside a tunnel lining is preliminarily determined; Step S2, a flexible loading grid is covered on the load-acting road surface of the tunnel scale model, and the output load of each modularized magnetic grid unit is calibrated; Step S3, the preliminarily determined load value is converted into an electromagnetic pulse parameter value, a plurality of electromagnetic pulse signals are controlled to be generated and sent to the micro electromagnetic vibration actuators in the grid units, so as to simulate the load; Step S4, actual tunnel lining strain data are collected in real time, an optimal electromagnetic pulse load distribution is back-calculated through an adaptive algorithm based on a digital twin model, and the application of the simulated load is continuously adjusted.

[0040] Further technical solutions, load simulation based on the load loading times of the tunnel scale model is performed, strain data after the simulated load is applied are collected, and lining fatigue damage is calculated according to the collected data.

[0041] Among them, the dynamic load applied to the road surface inside the tunnel lining includes road traffic load and subway traffic load, and the damage evaluation method proposed in the embodiment is introduced in more detail through the following two examples.

[0042] First, the dynamic moving load of the simulated highway tunnel inside road is taken as an embodiment, and the tunnel lining damage evaluation method includes: In step S1, the parameters of the tunnel scale model are designed based on similarity theory. Specifically, the geometric similarity ratio , the gravitational acceleration and the density similarity ratio are taken as the basic similarity ratios, the similarity relationship is obtained based on the similarity theory, and Table 1 is shown as follows.

[0043] Table 1 Similarity ratio of each parameter of the tunnel

[0044] Further, a tunnel scale model is built. The size of the model is: the diameter of the prototype tunnel is 10 m, the diameter of the model tunnel lining structure is 50 cm; the length of the model tunnel lining structure is 1 m; the internal road of the prototype tunnel is 6 m, and the model tunnel lining structure is 30 cm. In addition, the material similarity of the model is: the model lining structure adopts C50 concrete (elastic modulus = 34.50 GPa), and the target material elastic modulus of the scale model is 1.73 GPa; the model road structure adopts C30 concrete (elastic modulus = 30.00 GPa), and the target material elastic modulus of the scale model is 1.50 GPa.

[0045] Finally, the load is calculated. The load on the internal road of the tunnel lining includes the overburden load and the traffic load. The calculation of the overburden load is: the overburden pressure is 570.10 kPa when the tunnel depth is 30 m, and the overburden pressure is 20.36 kPa after similarity ratio conversion.

[0046] The calculation of the traffic load is: in this embodiment, two-axle passenger cars, low-speed trucks and small passenger cars are taken as the traffic load simulation objects, the maximum allowable total mass limit of the above three types of vehicles is summarized according to “Automobile, Trailer and Automobile Train External Outline Size, Axle Load and Mass Limit” (GB 1589-2016); at the same time, the road grade of the tunnel engineering in this embodiment is specified as urban trunk road, 6-lane two-way, double-deck, the design load is urban A-level vehicle, the standard load truck of urban A-level vehicle load adopts five-axle truck loading, the total weight is 700 kN; according to existing research, it is generally believed that the pulse repetition frequency of traffic load (i.e. the loading pulse repetition frequency) is 1 Hz ~5Hz; in addition, the relationship between driving speed and loading time is determined as: In the above formula, is the vehicle length.

[0047] Based on the above design, the loading parameters of dynamic loads of different vehicle types can be calculated and determined, which are shown in the following table 2.

[0048] Table 2 Loading parameters of dynamic loads of different vehicle types

[0049] In step S2, a flexible loading grid is covered on the surface of the internal road of the tunnel scale model under the action of the load, and the output load of each modular magnetic grid unit is calibrated. Specifically, a 30x100 cm magnetic flexible loading grid is laid on the surface of the model road, covering the vehicle action area, each grid unit in the grid is linked through a Halbach permanent magnet array, and the grid is fixed on the road surface using high-strength welding glue to avoid grid slipping during loading, so as to complete the arrangement of the modular grid unit.​

[0050] Further, in order to ensure the accuracy of the final load loading, the output load of each modular magnetic attraction grid unit is calibrated. That is, the 4x4 electromagnetic vibration actuator array of each grid unit works cooperatively, uses the acceleration sensor and the pressure sensor to monitor the vibration load data in real time, and adjusts the electromagnetic pulse load distribution adaptively according to the set load value and the monitored vibration load data, controls the generation of multiple electromagnetic pulse signals. By calibrating the pulse peak value and the pulse repetition frequency through the acceleration sensor (accuracy ±0.01g), it is ensured that the maximum dynamic load of a single grid unit meets the model wheel pressure impact requirements.

[0051] In step S3, the preliminary determined load value is converted into electromagnetic pulse parameter value, and multiple electromagnetic pulse signals are generated and sent to the micro electromagnetic vibration actuators in each grid unit to simulate the load. Specifically, for the simulation of the above three typical vehicle types (corresponding to the model load amplitude of 3.31N, 5.51N, 22.05N), the load is converted into electromagnetic pulse parameters through the intelligent control module, which can be expressed as: ; Among them, is the pulse current intensity (A), is the vibration load (N), is the actuator force-electricity conversion coefficient, is the pulse width ( ). Therefore, 3.31N, 5.51N, 22.05N correspond to 0.033A, 0.037A, 0.110A respectively, and the vibration frequency is set as the model pulse repetition frequency shown in Table 2 above.

[0052] As an embodiment, the fiber Bragg grating sensor is pasted on the inner surface of the tunnel lining structure, such as the left and right arch feet, the left and right arch waist, the vault, etc., to collect the vertical compressive stress , and verify whether the simulation conforms to the dynamic response formula under the vibration load, which is: ; Among them, is the maximum strain amplitude, is the model vibration frequency.

[0053] In step S4, the actual tunnel lining strain data is collected in real time, and the optimal electromagnetic pulse load distribution is deduced through the adaptive algorithm based on the digital twin model, and the application of the simulated load is continuously adjusted to effectively compensate for the difference between the load simulation and the actual situation, thereby realizing the accurate simulation of complex dynamic load and laying a foundation for the subsequent accurate damage assessment.

[0054] As an implementation, on the basis of the load simulation described above, the load loading times N = 10 are set 4 times, through the cyclic loading simulation of repeated vehicle passing, according to the real-time collected pavement surface strain (sampling rate 1kHz), based on the Miner fatigue accumulation rule, the damage degree is calculated, which is: ; is the number of cycles of the i level pulse load, is the fatigue life under the load, the value is related to the lining strain data, and the relationship can be obtained through material test, for example, the empirical formula of gypsum material is derived as: , wherein A and B are both constants related to the characteristics of gypsum material, A reflects the inherent fatigue performance of the material, and B represents the sensitivity of strain to fatigue life; preferably, when ε = 500 , = 5 × 10 4 times.

[0055] Finally, according to the damage degree, it is judged whether the tunnel lining structure is damaged or not, such as when the visible rut (depth ≥ 2mm) appears on the road surface, it indicates that the tunnel lining structure is damaged.

[0056] Example three The embodiment proposes a flexible mesh tunnel lining damage evaluation method, which is realized based on the flexible mesh tunnel lining multi-scale load simulation system proposed in example one. In this embodiment, taking the simulation of subway tunnel traffic dynamic load as an example, the tunnel lining damage evaluation method thereof includes: In step S1, dynamic load parameter conversion is performed. Specifically, taking the "Qingdao Sifang" B1 type rail transit subway vehicle as an example, the subway vehicle of this type adopts 6-vehicle marshalling, has 6 carriages, and the length is 19m. The loading length after scaling is 0.95m; the highest operating speed is 80km / h, and the scaled speed is 98.39m / s; the load amplitude of the train under normal operating conditions is 6900kN, and the scaled amplitude is 826.5N; the pulse repetition frequency is 2.75Hz, and the corresponding loading frequency after scaling is 12.30Hz; the pulse width is set to 300 .

[0057] ​In step S2, the grid arrangement is performed. Specifically, for the subway track area (double track, track gauge 1.435 m, after scaling 7.175 cm), a 100x10 cm magnetic attraction grid is laid on the model track surface to cover the track and ballast area, and a single grid unit is spliced by a Halbach permanent magnet array. The track contact part of the grid is coated with a non-slip coating (friction coefficient ≥ 0.8) to prevent the grid from slipping during train load simulation.

[0058] Further, the electromagnetic vibration actuators are calibrated in coordination, i.e., the subway load has the characteristics of continuous periodicity, and the pulse timing difference (≤5 ms) of adjacent grid units needs to be calibrated. Therefore, the electromagnetic pulse delay time of the 4x4 electromagnetic vibration actuator array is adjusted by the distributed control module to ensure that the propagation speed of the load along the track direction matches the scaled speed (98.39 m / s) of the subway vehicle, simulating the continuous rolling effect of the train wheelset. The calculation formula of the delay time is: ; wherein, is the pulse delay time (ms) of adjacent grids, is the grid unit spacing (100 mm), is the scaled train speed (98.39 m / s).

[0059] In step S3, according to the parameters in step S1: load amplitude 826.5 N, frequency 12.30 Hz, etc., the pulse current intensity is also converted to 0.276 A, and the dynamic load is applied in this way.

[0060] As an embodiment, the optical fiber grating sensors are arranged on the tunnel lining track foundation, arch waist and ballast bottom to collect vertical compressive stress, horizontal displacement and vibration acceleration, and the real-time current and frequency parameters of the grid unit are recorded synchronously to verify whether the dynamic load of the above loading meets the subway load dynamics response formula, which is: ; wherein, is the maximum compressive stress amplitude, is the phase difference of adjacent grids.

[0061] In step S4, the actual tunnel lining strain data is collected in real time, and the optimal electromagnetic pulse load distribution is deduced by the adaptive algorithm based on the digital twin model, and the application of the simulated load is continuously adjusted to effectively compensate for the differences between the load simulation and the actual situation, thereby realizing the accurate simulation of complex dynamic loads and laying a foundation for subsequent accurate damage assessment.

[0062] As an embodiment, fatigue damage analysis is carried out based on the above load simulation scheme. Specifically, according to the cumulative effect of 300 train trips per day and 10 years of operation, the loading frequency N = 10 5 times is set, and the track bed strain force is collected at a sampling rate of 2 kHz. On this basis, based on the pre-trained deep neural network model reflecting the relationship between electromagnetic pulse load-lining strain-crack propagation, the vibration frequency, amplitude and strain data are input, and the crack propagation rate is output. Further, the damage degree can be further measured by the output crack propagation rate and crack length.

[0063] As an embodiment, the model is designed based on the mapping relationship between electromagnetic pulse load parameters-lining strain-crack propagation rate, adopts a multi-layer perception (MLP) architecture, and mainly learns the complex correlation between input features and output targets through nonlinear transformation. The training process of the model is as follows: First, through tunnel scale model test, electromagnetic pulse load with different parameters (frequency, amplitude) is applied, and lining strain (fiber Bragg grating sensor) and crack propagation process (high-precision image measurement) are synchronously collected. Load parameter-strain-crack propagation rate sample data is calculated to build a data set. Preferably, finite element software is used to simulate the lining mechanical response and crack propagation under different load conditions to supplement the sample size of experimental data.

[0064] Secondly, the data in the data set is cleaned: abnormal values (such as strain mutation data caused by sensor failure) are removed, and normalized: all features (frequency, amplitude, strain) and target values (crack propagation rate) are mapped to the [0, 1] interval to eliminate the interference of dimension difference on model training. After that, it is divided into training set (model learning), validation set (parameter adjustment) and test set (final evaluation) according to the ratio of 7:1:2.

[0065] Then, the network architecture is built. The network model includes: Input layer: including 3 neurons, corresponding to 3 input features: vibration frequency (Hz) of electromagnetic pulse, load amplitude (kN), and strain data of lining ).

[0066] Hidden layer: contains 3 fully connected layers, with neuron numbers of 128→64→32 (decreasing layer by layer); ReLU function (rectified linear unit) is used in the activation layer to introduce nonlinear mapping ability, solve the problem that linear model cannot fit complex mechanical relationship, and avoid gradient disappearance phenomenon; regularization mechanism: Dropout layer (dropout rate 20%) is added after the first two hidden layers to prevent the model from over-relying on specific features (such as data in a certain frequency interval) and reduce the risk of overfitting.

[0067] The output layer includes one neuron with a linear activation function, directly predicting the crack propagation rate of the lining (unit: mm / cycle loading times) without the need for nonlinear transformation to reflect the numerical size of the true physical quantity.

[0068] Finally, using the dataset, the network parameters (weights and biases) are iteratively adjusted to minimize the error between the predicted crack propagation rate and the true value. By calculating the root mean square error (RMSE) and the mean absolute error (MAE) of the test set, the prediction accuracy is evaluated and the model is tested with new scene data (such as different lining materials, load types) that did not participate in the training, ensuring its reliable prediction ability in actual engineering, and completing the training of the model.

[0069] Through the above scheme of the embodiment, accurate simulation of multi-scale load of tunnel lining can be realized, laying a foundation for subsequent accurate damage analysis.

[0070] The steps and methods involved in the above embodiments two and three correspond to embodiment one, and the specific implementation can refer to the relevant description part of embodiment one.

[0071] Those skilled in the art should understand that each module or step of the present application described above can be realized by a general computer device, alternatively, they can be realized by program codes executable by a computing device, so that they can be stored in a storage device and executed by a computing device, or they can be respectively manufactured into each integrated circuit module, or a plurality of modules or steps among them can be manufactured into a single integrated circuit module to realize. The present application is not limited to any specific combination of hardware and software.

[0072] The above only describes the preferred embodiments of the present application, and the specific implementation of the present application is described in conjunction with the drawings, but it is not a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications or variations made on the basis of the technical solutions of the present application without creative labor are still within the protection scope of the present application.

Claims

1. A flexible mesh based multi-scale load simulation system for tunnel lining, characterized in that, The system comprises: a flexible loading grid for covering the load-acting road surface of the tunnel scale model, comprising a plurality of modular magnetic grid units; an array of micro electromagnetic vibration actuators integrated inside each grid unit, each micro electromagnetic vibration actuator in the array acting electromagnetic pulse signals to the internal coil to drive mechanical vibration; a pulse generator for generating electromagnetic pulse signals and sending them to each micro electromagnetic vibration actuator; an intelligent control module for converting the set load value into electromagnetic pulse parameter value and controlling the generation of multiple electromagnetic pulse signals.

2. The flexible mesh-based tunnel lining multiscale load simulation system of claim 1, wherein, It also includes a signal acquisition unit; The signal acquisition unit includes a micro acceleration sensor and a pressure sensor embedded in the bottom of each grid unit for real-time monitoring of the vibration load data applied in the grid unit; wherein the vibration load data includes the pulse peak value, the pulse repetition frequency and the pulse pressure peak value acting on the internal road surface of the lining. The intelligent control module is also used to adaptively adjust the electromagnetic pulse load distribution according to the set load value and the monitored vibration load data, and control the generation of multiple electromagnetic pulse signals.

3. The flexible mesh-based tunnel lining multiscale load simulation system of claim 2, wherein, The electromagnetic pulse load distribution includes the pulse peak value, the repetition frequency, the pulse width, the pulse timing and the duration of the electromagnetic pulse signal applied to each grid unit.

4. The flexible mesh-based tunnel lining multiscale load simulation system of claim 1, wherein, Each modular magnetic grid unit is made of silica gel-aramid composite material, and a permanent magnet array is embedded in the periphery of each grid unit for magnetic assembly of multiple modular magnetic grid units to form a flexible loading grid.

5. The flexible mesh-based tunnel lining multiscale load simulation system of claim 1, wherein, The periphery of each modular magnetic grid unit is uniformly provided with a silica gel damping vibration isolation ring.

6. The flexible mesh-based tunnel lining multiscale load simulation system of claim 2, wherein, The signal acquisition unit also includes a fiber Bragg grating sensor arranged on the surface of the tunnel lining for real-time acquisition of actual tunnel lining strain data; The intelligent control module is also used to back-calculate the optimal electromagnetic pulse load distribution through the pre-loaded adaptive algorithm based on the digital twin model according to the acquired actual lining strain data, to regulate the generation of multiple electromagnetic pulse signals.

7. The flexible mesh-based tunnel lining multiscale load simulation system of claim 6, wherein, According to the acquired actual lining strain data, the optimal electromagnetic pulse load distribution is back-calculated through the pre-loaded adaptive algorithm based on the digital twin model, including: constructing a digital twin model of the tunnel lining and its load-acting environment; inputting the real-time acquired actual lining strain data into the digital twin model of the tunnel lining, comparing the target strain corresponding to the lining under different electromagnetic pulse load distributions preset in the digital twin model with the actual strain, and judging whether to adjust the electromagnetic pulse load parameters according to the comparison result; if the judgment is yes, the adaptive algorithm is used to make the actual strain close to the target strain as the optimization goal, and the required adjusted electromagnetic pulse load parameters are calculated, i.e. the back-calculated optimal electromagnetic pulse load distribution.

8. A flexible mesh based multi-scale load simulation method for tunnel lining, characterized in that, The flexible grid-based tunnel lining multi-scale load simulation system according to any one of claims 1-7 comprises: designing model parameters based on similarity theory, building a tunnel scale model, and preliminarily determining the load value applied to the internal road surface of the tunnel lining; covering the flexible loading grid on the load-acting road surface of the tunnel scale model, and calibrating the output load of each modular magnetic grid unit; The preliminary determined load value is converted into electromagnetic pulse parameter value, a plurality of electromagnetic pulse signals are controlled to be generated and sent to the micro electromagnetic vibration actuators in each grid unit to simulate the load; Real-time acquisition of actual tunnel lining strain data, through the adaptive algorithm based on the digital twin model, the optimal electromagnetic pulse load distribution is back calculated, and the application of the simulated load is continuously adjusted.

9. The flexible mesh-based tunnel lining multiscale load simulation method of claim 8, wherein, The calibration of the load output by each modular magnetic attraction grid unit is: Real-time monitoring of vibration load data; wherein the vibration load data includes the pulse peak value, the pulse repetition frequency of the vibration load, and the pulse pressure peak value acting on the internal road surface of the lining; According to the set load value and the monitored vibration load data, the electromagnetic pulse load distribution is adaptively adjusted, and the generation of a plurality of electromagnetic pulse signals is controlled; wherein the electromagnetic pulse load distribution includes the pulse peak value, the repetition frequency, the pulse width, the pulse timing and the duration of the electromagnetic pulse signal applied to each grid unit.

10. The flexible mesh-based tunnel lining multiscale load simulation method of claim 8, wherein, Based on the tunnel scale model, the load simulation of the load loading times is set, the strain data after the application of the simulated load is collected, and the lining fatigue damage is calculated according to the collected data.

Images