Aftershock prediction method and system based on earthquake information and seismic data

Abstract

The application discloses a kind of across fault tunnel lining metastable state identification and regulation method, specific steps are: input test model information and sensor measuring point information;Establish static and small earthquake dynamic benchmark;Collect time history of multiple sensors in hierarchical loading, calculate four types of comprehensive indexes of dynamic degradation, deformation development, soil pressure redistribution and structural strain damage;Determine spatial correlation correction coefficient and trend persistence correction coefficient according to the spatial concentration degree and continuous deterioration trend of abnormal response, calculate the metastable state comprehensive discriminant index by weighting, combined with multi-index collaborative triggering, online determination of six evolution levels from stable state to near instability;Finally, grade output hierarchical regulation, including improving sampling frequency, inserting sweep frequency retest, reducing amplitude or terminating test.The application can capture the precursor of damage in real time, improve data integrity and security.

Description

Technical Field

[0001] This invention relates to the field of tunnel engineering model testing and seismic testing of underground structures, and in particular to a method for identifying and controlling the metastable state of tunnel lining across faults. This method is applicable to the identification of precursory failures of the lining-surrounding rock system, the capture of key responses, and the feedback adjustment of the test loading process during graded seismic loading. Background Technology

[0002] Under seismic loading, tunnels crossing faults are simultaneously affected by ground inertia, fault slippage, deformation of weak interlayers, and changes in the interaction between the lining and surrounding rock. Fault fracture zones and their adjacent areas are prone to concentrated slippage deformation, redistribution of earth pressure, local strain accumulation, and degradation of dynamic characteristics. These responses typically do not directly progress from a stable state to overall failure, but rather undergo a metastable stage of gradual damage accumulation approaching instability. For shaking table tests, if this stage cannot be identified before the lining develops through cracks, local collapses, or the load-bearing system fails, it is easy to miss critical responses prior to failure, and it is also detrimental to adjusting subsequent loading levels and sampling strategies.

[0003] Existing technologies have made relevant improvements. For example, Chinese patent application CN106226808A discloses an experimental device for simulating the seismic response of tunnels under fault displacement, and Chinese patent application CN102645310A discloses a shaking table test method for soft soil iron structures. These existing technologies provide a technical foundation for fault displacement simulation, boundary condition setting, and response measurement. However, their focus is mostly on the construction of the experimental device, the method of realizing fault displacement, or the arrangement of sensors, with insufficient attention paid to the metastable stage of the lining developing from a stable response to local instability during graded loading. Precursors of failure often manifest as the co-evolution of multiple responses (such as increased relative displacement, development of differential soil pressure between the hanging wall and footwall, accumulation of residual strain, and decrease in fundamental frequency), making it difficult to determine the model state in real time by relying solely on a single measurement point threshold or offline analysis.

[0004] Furthermore, existing shaking table tests typically involve loading at predetermined levels, with sampling frequencies, retesting steps, and subsequent loading schemes often fixed in advance. This lacks a feedback mechanism to dynamically adjust sampling, retesting, and loading based on metastable state identification results, easily leading to insufficient data collection before failure or causing the model to rapidly skip metastable states and directly enter instability. Therefore, there is an urgent need to propose a metastable state collaborative identification and feedback control method for shaking table tests of cross-fault tunnel linings to improve the timeliness of capturing failure precursors and the completeness of test data. Summary of the Invention

[0005] To address the aforementioned technical shortcomings, the purpose of this invention is to provide a method for identifying and controlling the metastable state of lining in cross-fault tunnels. This method can solve the problem that existing cross-fault tunnel shaking table tests mainly rely on preset loading levels, on-site observation, and post-test data analysis, making it difficult to identify the metastable stage of the lining as it develops from a stable response to local instability.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0007] This invention provides a method for identifying and controlling metastable states in cross-fault tunnel linings, comprising the following steps:

[0008] S1: Obtain the parameters of the scaled model of the tunnel lining across the fault, the layout parameters of the fault fracture zone and / or weak interlayer, and the sensor configuration information. Based on the location of the fault plane, the fault fracture zone and / or weak interlayer, divide the scaled model of the tunnel lining across the fault into multiple model partitions. Assign each sensor to the corresponding model partition and lining part according to the measured coordinates to form the correspondence between the measuring points and the model partition and lining part.

[0009] S2: Before the shaking table is formally graded and loaded, static reference data and low-amplitude dynamic response data are collected to determine the initial state parameters of the scaled model of the cross-fault tunnel lining.

[0010] S3: During the graded loading process of the shaking table, the time history data of each sensor are collected simultaneously, and the comprehensive index of dynamic degradation, the comprehensive index of deformation development, the comprehensive index of earth pressure redistribution, and the comprehensive index of structural strain damage are calculated.

[0011] S4: Based on the measurement point correspondence established in step S1, determine the model partition and lining part where the abnormal response is located, and determine the spatial correlation correction coefficient; based on the development of the abnormal response in continuous loading levels or continuous time windows, determine the trend persistence correction coefficient.

[0012] S5: Based on the comprehensive index of dynamic degradation, the comprehensive index of deformation development, the comprehensive index of earth pressure redistribution, the comprehensive index of structural strain damage, the spatial correlation correction coefficient, and the trend persistence correction coefficient, calculate the metastable comprehensive discrimination index, and combine it with the multi-index collaborative triggering rule to determine the metastable evolution level of the scaled model of the cross-fault tunnel lining online.

[0013] S6: Based on the metastable evolution level, output feedback control measures to the vibration table control system, data acquisition system, or test monitoring terminal.

[0014] Preferably, the model partitions include the affected section of the fault hanging wall, the affected section of the fault footwall, the affected section of the fault fracture zone and / or the affected section of the weak interlayer, and the standard section far from the fault; the lining parts include the arch crown, arch waist, sidewalls and invert arch; the sensors include accelerometers, displacement gauges, earth pressure cells and strain gauges.

[0015] Preferably, the comprehensive deformation development index includes a relative displacement deformation index determined by displacement measurement data on both sides of the fault fracture zone and / or weak interlayer; when the relative displacement increment within adjacent loading levels or adjacent time windows exceeds a set threshold, it is determined that there is an accelerated development of displacement deformation in the lining adjacent to the fault.

[0016] Preferably, the comprehensive index for earth pressure redistribution includes an index for the development of the difference in earth pressure between the hanging wall and the footwall, determined by earth pressure measurement data from the hanging wall and footwall affected by the fault. When the representative earth pressure difference between the hanging wall and the footwall continues to increase, the ratio of earth pressure on both sides exceeds a set threshold, or the earth pressure on one side decreases while the earth pressure in adjacent areas increases, it is determined that the earth pressure around the lining is developing in a biased manner or that there is a phenomenon of contact pressure unloading-transfer.

[0017] Preferably, the comprehensive index of dynamic degradation is determined by one or more of the following: change in acceleration amplification factor, reduction in fundamental frequency, or change in damping ratio; the comprehensive index of structural strain damage is determined by one or more of the following: peak strain utilization rate, residual strain, or strain increment.

[0018] Preferably, the spatial correlation correction coefficient is determined based on the degree of concentration of anomalous responses within the same or adjacent model partitions. When two or more of the anomalous indicators among the comprehensive indicators of dynamic degradation, deformation development, earth pressure redistribution, and structural strain damage appear in the same or adjacent model partitions, it is determined that there is a spatial correlation anomaly in the local area corresponding to the model partition or the combination of adjacent model partitions, and spatial correlation correction is triggered to increase the spatial correlation correction coefficient. When the anomalies are dispersed and do not form local clusters, spatial correlation correction enhancement is not triggered.

[0019] Preferably, the trend persistence correction coefficient is determined based on the development of the abnormal response within consecutive loading levels or consecutive time windows. When at least one of the following indicators—dynamic degradation comprehensive index, deformation development comprehensive index, earth pressure redistribution comprehensive index, and structural strain damage comprehensive index—or the metastable comprehensive discrimination index, shows continuous or accelerated growth within multiple consecutive loading levels or multiple consecutive real-time calculation windows, and the growth rate or magnitude exceeds the corresponding threshold, the abnormal response is determined to have a trend persistence characteristic, and the trend persistence correction coefficient is increased.

[0020] Preferably, the metastable comprehensive discrimination index is determined according to the following formula:

[0021] ;

[0022] in, It is a metastable comprehensive discrimination index; This is a spatial correlation correction coefficient. For trend continuation correction coefficient, As a comprehensive indicator of power degradation, As a comprehensive indicator of deformation development, It is a comprehensive index for the redistribution of earth pressure. It is a comprehensive index of structural strain and damage. , , and For the corresponding weights, and .

[0023] Preferably, the metastable evolution levels include a stable state, a damage initiation state, a slightly metastable state, a moderately metastable state, a highly metastable state, and a near-instability / destruction state.

[0024] Preferably, the feedback control measures include:

[0025] When the damage is in the nascent stage or in a mild metastable state, increase the sampling frequency during critical response periods;

[0026] When in a moderately metastable state, insert a low-amplitude sweep frequency retest;

[0027] When in a highly metastable state, reduce the peak ground acceleration of the input seismic wave for the next loading level or suspend loading checks;

[0028] When the system is in a state close to instability / failure, the loading process is terminated and key response data before and after the failure is saved.

[0029] Beneficial effects:

[0030] 1. Breaking through the limitations of traditional experiments that rely on post-event offline analysis or post-damage review, this study shifts the monitoring focus to the metastable stage before structural instability, enabling real-time capture of the critical characteristics of the lining's evolution from stable response to local instability, thus providing methodological support for damage early warning.

[0031] 2. Taking full account of the discontinuous deformation effect caused by fault slippage, two core indicators were specifically extracted: "relative slippage deformation on both sides of the fault" and "differential development of earth pressure on the hanging wall / footwall". This overcomes the problem of slow response of traditional single-point absolute displacement and absolute stress at the fault location, and significantly improves the characterization accuracy of the abnormal interaction between the surrounding rock and the lining near the fault slippage zone.

[0032] 3. It integrates four types of responses: dynamic characteristics, deformation, earth pressure, and strain, and introduces correction coefficients in two dimensions: spatial correlation correction (focusing on fault vulnerable areas) and trend persistence correction (focusing on continuous deterioration of response). This effectively filters out misjudgments caused by poor local contact of the sensor or single occasional vibration, ensuring the reliability of the discrimination index.

[0033] 4. The metastable identification results are fed back to the vibration table control system and data acquisition system in real time, and the sampling frequency is dynamically adjusted, frequency sweep retest is inserted, subsequent loading amplitude is controlled, and even the test is terminated. This not only protects the high-cost test model from sudden damage, but also ensures the complete acquisition of high-value mechanical data before damage, which greatly improves the test efficiency and data validity. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art 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.

[0035] Figure 1 A flowchart illustrating a method for identifying and controlling metastable states in cross-fault tunnel lining, provided in an embodiment of the present invention;

[0036] Figure 2 This is a scaled-down model of a cross-fault tunnel lining and a schematic diagram of the measuring point layout provided in an embodiment of the present invention. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] See Figure 1 This embodiment provides a method for metastable state identification and control of cross-fault tunnel lining, including the following steps:

[0039] S1. Input the experimental model information and sensor measurement point information;

[0040] In this embodiment, a scaled-down model of the tunnel lining across a fault (hereinafter referred to as the model) is arranged inside a shaking table model box. The model box is filled with materials similar to surrounding rock or soil, and fault fracture zones, weak interlayers, or soft-hard soil interfaces are set. First, the geometric parameters, material parameters, fault fracture zone and / or weak interlayer arrangement parameters, and configuration information of each sensor of the model are obtained.

[0041] Based on the spatial location of the fault plane or weak interlayer in the model, the model is divided longitudinally along the tunnel into the fault hanging wall influence section, the fault footwall influence section, the fault fracture zone and / or weak interlayer influence section, and the standard section far from the fault. The standard section far from the fault is the basic monitoring area, and the rest are key monitoring areas.

[0042] Within each zone, the lining components such as the arch crown, arch waist, sidewalls, and invert are further determined based on the location of the lining cross-section. Accelerometers can be deployed on the platform, surrounding rock or soil, lining arch crown, and areas adjacent to the fault; displacement gauges can be deployed on both sides of the fault fracture zone, arch crown, sidewalls, or near the invert to monitor overall deformation, cross-sectional convergence, and relative displacement on both sides of the fault; earth pressure cells can be deployed on the outer arch waist, sidewalls, invert, and areas adjacent to the fault to monitor changes in lining-surrounding rock contact pressure; strain gauges can be deployed on the inner and outer surfaces of the lining, arch waist, sidewalls, invert, and vulnerable areas adjacent to the fault to monitor the development of local tensile and compressive strain.

[0043] In this embodiment, within each zone, each monitoring section is further divided into an arch top zone, a left arch waist zone, a right arch waist zone, a left wall zone, and a right wall zone along the lining section direction; wherein, the left wall zone and the right wall zone are symmetrically located at the side wall parts on both sides of the lining, the left arch waist zone and the right arch waist zone are symmetrically located at the arch waist parts on both sides of the lining, and the arch top zone is located at the center of the top of the lining.

[0044] Within different cross-sectional zones, different types of sensors are configured according to monitoring requirements, such as... Figure 2 As shown:

[0045] Accelerometers, displacement gauges, earth pressure cells, and strain gauges are installed in the left and right wall areas respectively to synchronously monitor the acceleration response, relative or absolute displacement, lining-surrounding rock contact pressure, and local tensile and compressive strain at these locations.

[0046] Accelerometers, earth pressure cells, and strain gauges are installed in the left arch waist area, right arch waist area, and arch crown area, respectively, to monitor the acceleration response, lining-surrounding rock contact pressure, and local tensile and compressive strain in these areas.

[0047] Before the experiment began, the number, type, coordinates, sampling frequency, installation direction, and model zone of each measuring point were entered into the data acquisition system to establish a correspondence between the measuring points and the model zones and lining parts. This correspondence is used to subsequently determine whether abnormal responses are concentrated in fault fracture zones, weak interlayer influence sections, or adjacent lining parts.

[0048] S2. Establish static and small-scale seismic load data. Before formal graded seismic loading, collect static load data, including initial displacement at each displacement measuring point, initial earth pressure at each earth pressure measuring point, and initial strain at each strain measuring point. Static load data is used for subsequent zero-point correction, residual displacement calculation, residual strain calculation, and contact pressure variation analysis.

[0049] Subsequently, low-amplitude seismic waves or low-amplitude swept-frequency signals are input to acquire the model's small-amplitude dynamic response, obtaining the initial acceleration amplification factor, initial fundamental frequency, and initial damping ratio. The low-amplitude dynamic response is not intended to cause lining damage, but rather to determine the model's dynamic characteristics in a basically intact state. During the formal loading process, all levels of response are compared to this baseline state.

[0050] S3. Data Acquisition and Preprocessing: Acceleration, displacement, earth pressure, and strain data are acquired in real time during the graded loading process on the shaking table. Acceleration includes the input acceleration from the table surface, the acceleration of the surrounding rock or soil, and the lining acceleration. For each loading stage or each calculation time window, data preprocessing is performed on each channel, including time synchronization, zero-point correction, filtering and noise reduction, and outlier removal. After each loading stage, the comprehensive indexes of dynamic degradation, deformation development, earth pressure redistribution, and structural strain damage are calculated.

[0051] The dynamic degradation comprehensive index is used to reflect changes in the dynamic performance of the lining-surrounding rock system, and is determined by changes in the acceleration amplification factor, the decrease in the fundamental frequency, and the change in the damping ratio. When the acceleration amplification factor is significantly increased or decreased compared to the reference state, and the fundamental frequency shows a continuous decrease, it indicates that the model stiffness, contact state, or degree of local damage has changed.

[0052] The comprehensive deformation development index is used to reflect the cumulative deformation of the lining and the area adjacent to the fault, including peak displacement, residual displacement, residual displacement increment, and cross-sectional convergence. For cross-fault tests, the focus is on extracting the relative displacement between displacement measuring points on both sides of the fault fracture zone and / or weak interlayer, i.e., the relative displacement deformation index. When the relative displacement increment between adjacent loading levels or within adjacent time windows exceeds a set threshold, it is determined that there is an accelerated development of displacement deformation in the lining adjacent to the fault.

[0053] The comprehensive index for earth pressure redistribution reflects changes in the contact pressure between the lining and the surrounding rock, including changes in earth pressure, residual earth pressure, and the development index of the difference in earth pressure between the hanging wall and footwall. The development index of the difference in earth pressure between the hanging wall and footwall is determined based on the responses of corresponding earth pressure measuring points in the affected sections of the fault's hanging wall and footwall. When the representative earth pressure difference between the hanging wall and footwall continuously increases, the ratio of earth pressure on both sides exceeds a set threshold, or the earth pressure on one side decreases while the earth pressure in adjacent areas increases, it is determined that there is a biased development of earth pressure around the lining or a phenomenon of contact pressure unloading-transfer.

[0054] The comprehensive structural strain-damage index is used to reflect the development of local damage and is determined by peak strain utilization, residual strain, and strain increment. When the strain in the arch waist, sidewall, or invert near the fault continues to increase, accompanied by the accumulation of residual strain, it indicates that the lining in that area may have entered the damage development stage.

[0055] The above-mentioned indicators are dimensionless based on the results of preliminary tests, similarity ratios, material strengths, and numerical simulations, so that indicators with different dimensions can be compared under the same discrimination framework.

[0056] S4. Calculation of spatial correlation correction coefficient and trend persistence correction coefficient; Since a single measurement point anomaly may be caused by local contact, sensor disturbance or accidental response, this implementation does not directly use the peak value of a single index as the metastable criterion, but combines the spatial location and development trend of the abnormal response for correction, that is, the identification result is corrected according to the partition location and development trend.

[0057] The spatial correlation correction coefficient is determined based on the degree of concentration of anomalous responses within the same or adjacent model partitions.

[0058] The specific judgment method is as follows: Based on the correspondence of measuring points established in step S1, determine the model partition and lining location of each abnormal response; when two or more of the comprehensive indicators of dynamic degradation, deformation development, earth pressure redistribution, and structural strain damage appear concentrated in the same model partition or adjacent model partitions, such as simultaneously concentrated in the arch crown area, arch waist area, sidewall area, or invert arch area, it is determined that the lining location has a risk of local weakness or concentrated damage, and spatial correlation correction is triggered; if different types of anomalies appear scattered in different model partitions and do not form local spatial clusters, their spatial correlation is determined to be weak, and spatial correlation correction enhancement is not triggered. Anomaly types include dynamic degradation anomalies, deformation development anomalies, earth pressure anomalies, and strain damage anomalies.

[0059] The trend persistence correction coefficient is determined based on the development of the abnormal response within consecutive loading levels or consecutive time windows. For example, let k be the loading level number or real-time calculation window number, used to represent the current position of the data sequence participating in metastable state discrimination. If the shaking table test uses a graded loading method, then k represents the current k-th level of seismic motion loading; for example, levels 1, 2, and 3 correspond to k=1, k=2, and k=3, respectively. If the shaking table test uses a continuous real-time acquisition method, then k represents the k-th calculation window obtained according to a preset time length; for example, after dividing the continuously acquired data into 2-second, 5-second, or other preset time lengths, each data segment corresponds to a calculation window. k-1 represents the level or window preceding the current loading level or current calculation window, and k-2 represents the two levels or windows preceding the current loading level or current calculation window; correspondingly, k+1 and k+2 represent subsequent adjacent loading levels or calculation windows. By comparing the changes in the comprehensive indices of dynamic degradation, deformation development, earth pressure redistribution, and structural strain damage in continuous sequences such as k-2, k-1, k, k+1, and k+2, it can be determined whether the changes in these indices are random fluctuations, continuous growth, or accelerated development. When an index shows a peak only in a single k-sequence, and the preceding and following sequences do not show a continuous increasing trend, the system does not directly classify it as an abnormal trend. When an index continuously increases in multiple consecutive k-sequences, or its growth rate gradually increases, the system determines that the index has a continuous trend development characteristic and performs a trend-continuous correction on the metastable state identification results. This processing can reduce the impact of random fluctuations in a single loading cycle on the metastable state judgment.

[0060] S5. Metastable Co-identification and Classification; After completing the index calculation and correction, calculate the metastable comprehensive discrimination index according to the following formula:

[0061] ;

[0062] in, It is a metastable comprehensive discrimination index; This is a spatial correlation correction coefficient. For trend continuation correction coefficient, As a comprehensive indicator of power degradation, As a comprehensive indicator of deformation development, It is a comprehensive index for the redistribution of earth pressure. It is a comprehensive index of structural strain and damage. , , and For the corresponding weights, and .

[0063] Based on the metastable comprehensive discriminant index Based on the magnitude of the magnitude of the damage and the synergistic triggering of multiple indicators, the model state is divided into six levels: stable state, damage initiation state, slightly metastable state, moderately metastable state, highly metastable state, and near-instability / failure state. Among them, the stable state indicates that no significant anomalies are observed in any type of response, and the model is within the normal bearing capacity range; the damage initiation state indicates that local measuring points or a single type of indicator show anomalies, but a multi-indicator synergistic metastable characteristic has not yet formed; the slightly metastable state indicates that at least two types of indicators show synergistic anomalies; the moderately metastable state indicates that the abnormal response continues to develop near the fault; the highly metastable state indicates that the response values ​​corresponding to at least two types of indicators develop rapidly, accompanied by a significant trend of increase; and the near-instability / failure state indicates that the relative displacement deformation measured by representative displacement measuring points in the fault fracture zone and / or the lining near the weak interlayer, the peak strain or residual strain measured by strain measuring points in the same location, or the comprehensive index of earth pressure redistribution reach a preset threshold.

[0064] S6. Early warning level and test feedback control: After the metastable level is determined, the system will output the identification result to the vibration table control system, data acquisition system or test monitoring terminal for adjustment of the subsequent test process.

[0065] The output results should include at least the warning level (Level IV), triggering location (structural section), triggering measuring point (number / location), triggering index (Ms), triggering time (time / level), and recommended measures; the output format for the triggering location is model partition and lining location, and the output format for the triggering measuring point is the type of each sensor and the measuring point number.

[0066] Recommended measures include, but are not limited to, encrypted sampling, adjusting the level, pausing loading, and on-site inspection.

[0067] Feedback control measures are implemented according to the metastable state level:

[0068] When the model is in the damage initiation stage (corresponding to Level I warning) or mild metastable state (corresponding to Level II warning), increase the sampling frequency during the critical response period. Specifically, increase the sampling frequency of displacement, soil pressure and strain measurement points near the fault, and mark abnormal measurement points and abnormal locations.

[0069] When the model is in a moderately metastable state (corresponding to Level III warning), a low-amplitude frequency sweep retest is inserted to re-identify the current fundamental frequency and dynamic characteristics in order to obtain the changes in the dynamic parameters of the model under the current damage state.

[0070] When the model is in a highly metastable state (corresponding to Level IV warning), reduce the peak ground acceleration of the input seismic wave for the next loading level, or pause loading and check the cracks, contact status, and sensor status of the lining near the fault.

[0071] When the model is close to instability / destruction (corresponding to Level V warning), the loading is terminated and the key response data before and after the destruction is saved.

[0072] The aforementioned feedback control measures are implemented after confirmation by the test personnel, or they can be implemented automatically when the vibration table control system and data acquisition system meet safety conditions. For tests with different similarity ratios, different surrounding rock materials, and different fault arrangements, the sampling frequency, retest signal, peak acceleration adjustment amplitude, and state threshold are set based on the pre-test results or numerical simulation results.

[0073] To illustrate the specific calculation process of the method of the present invention, a scaled-down shaking table test of a cross-fault tunnel lining is used as an example. This example is only used to explain the implementation of the present invention and does not limit the specific values ​​of each index, weight, and threshold.

[0074] The test model included a weak interlayer traversing the tunnel lining. Displacement gauges, earth pressure cells, accelerometers, and strain gauges were placed near the weak interlayer. Before formal loading, the initial state of the model was obtained through static acquisition and low-amplitude frequency sweep measurement: the initial fundamental frequency was 18.5Hz, the initial relative displacement of the displacement measuring points on both sides of the fault was 0mm, the representative earth pressure difference between the hanging wall and the footwall was 3.2kPa, and the strain at the arch waist near the fault was near the initial zero point.

[0075] After the fourth level of loading, the system detected a certain degree of residual deformation in the vicinity of the fault fracture zone, an increased difference in earth pressure between the hanging wall and footwall, and residual accumulation of strain in the arch waist, but the decrease in the fundamental frequency was relatively small. After normalization, the comprehensive dynamic degradation index was obtained. =0.36, Comprehensive index of deformation development =0.48, comprehensive index of earth pressure redistribution =0.45, comprehensive structural strain-damage index =0.50. Since the anomalous response is mainly concentrated in the vicinity of the fault, the spatial correlation correction coefficient is taken. =1.10; Since some indicators have increased for two consecutive loading levels, the trend continuation correction coefficient is used. =1.05; , , and If we take the values ​​as 0.20, 0.30, 0.25, and 0.25 respectively, then:

[0076] ;

[0077] According to the preset grading rules, the system determines that the model is in a slightly metastable state. The output feedback control measures are: increase the sampling frequency of displacement, soil pressure and strain measurement points near the fault, and mark the key response period after the loading level.

[0078] After the fifth level of loading, the relative displacement between the two sides of the fault fracture zone increased from 2.1 mm in the fourth level to 4.8 mm, the representative earth pressure difference between the hanging wall and footwall increased from 11.5 kPa to 19.6 kPa, the strain utilization rate of the arch waist near the fault increased to 0.64, and the model fundamental frequency decreased to 15.8 Hz. After normalization, the following results were obtained: =0.52, =0.61, =0.58, =0.65. Since the deformation, earth pressure, and strain anomalies are concentrated in the fault fracture zone and its adjacent lining area, we take... =1.20; Due to the continuous increase in relative displacement deformation, soil pressure difference, and residual strain over multiple loading stages, take... =1.15; , , and If we still take the values ​​as 0.20, 0.30, 0.25, and 0.25 respectively, then:

[0079] ;

[0080] According to the preset grading rules, the system determines that the model has entered a highly metastable state. At this point, the model has not yet shown through cracks or overall instability, but the lining near the fault has already shown accelerated slip deformation, development of earth pressure bias, and accumulation of local strain. Based on this, the system outputs feedback control measures: inserting a low-amplitude frequency sweep retest to re-identify the current fundamental frequency, pausing loading and checking the cracks and sensor status at the arch waist and sidewalls near the fault, and simultaneously lowering the peak ground acceleration of the seismic wave to be input for the next loading level.

[0081] If, during subsequent loading or retesting, the relative displacement deformation measured by representative displacement measuring points of the fault fracture zone and / or the lining adjacent to the weak interlayer, the peak strain or residual strain measured by strain measuring points of the same location, or the comprehensive index of earth pressure redistribution reaches the preset threshold, the system determines that the model has entered a state close to instability / failure, and outputs feedback control measures to terminate loading and save key response data before and after failure.

[0082] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for identifying and controlling metastable states in cross-fault tunnel linings, characterized in that, include: S1: Obtain the parameters of the scaled model of the tunnel lining across the fault, the layout parameters of the fault fracture zone and / or weak interlayer, and the sensor configuration information. Based on the location of the fault plane, the fault fracture zone and / or weak interlayer, divide the scaled model of the tunnel lining across the fault into multiple model partitions. Assign each sensor to the corresponding model partition and lining part according to the measured coordinates to form the correspondence between the measuring points and the model partition and lining part. S2: Before the shaking table is formally graded and loaded, static reference data and low-amplitude dynamic response data are collected to determine the initial state parameters of the scaled model of the cross-fault tunnel lining. S3: During the graded loading process of the shaking table, time history data of each sensor are collected simultaneously, and the comprehensive index of dynamic degradation, the comprehensive index of deformation development, the comprehensive index of earth pressure redistribution, and the comprehensive index of structural strain damage are calculated. S4: Based on the measurement point correspondence established in step S1, determine the model partition and lining part where the abnormal response is located, and determine the spatial correlation correction coefficient; based on the development of the abnormal response in continuous loading levels or continuous time windows, determine the trend persistence correction coefficient. The spatial correlation correction coefficient is determined based on the degree of concentration of abnormal responses within the same or adjacent model partitions. When two or more of the following abnormal indicators appear in the same or adjacent model partitions: dynamic degradation comprehensive index, deformation development comprehensive index, earth pressure redistribution comprehensive index, and structural strain damage comprehensive index, it is determined that there is a spatial correlation anomaly in the local area corresponding to the model partition or the combination of adjacent model partitions, and spatial correlation correction is triggered to increase the spatial correlation correction coefficient. When the distribution is abnormally dispersed and no local clustering is formed, spatial correlation correction enhancement is not triggered; The trend persistence correction coefficient is determined based on the development of the abnormal response in successive loading levels or continuous time windows. When at least one of the following indices, namely the comprehensive index of dynamic degradation, the comprehensive index of deformation development, the comprehensive index of earth pressure redistribution, and the comprehensive index of structural strain damage, or the comprehensive index of metastable state, shows a continuous increase or accelerated increase within multiple consecutive loading levels or multiple consecutive real-time calculation windows, and the increase or growth rate exceeds the corresponding threshold, the abnormal response is determined to have a trend of continuous development, and the trend of continuous correction coefficient is increased. S5: Based on the comprehensive index of dynamic degradation, the comprehensive index of deformation development, the comprehensive index of earth pressure redistribution, the comprehensive index of structural strain damage, the spatial correlation correction coefficient, and the trend persistence correction coefficient, calculate the metastable comprehensive discrimination index, and combine it with the multi-index collaborative triggering rule to determine the metastable evolution level of the scaled model of the cross-fault tunnel lining online. The metastable comprehensive discrimination index is determined according to the following formula: ； in, It is a metastable comprehensive discrimination index; This is a spatial correlation correction coefficient. For trend continuation correction coefficient, As a comprehensive indicator of power degradation, As a comprehensive indicator of deformation development, It is a comprehensive index for the redistribution of earth pressure. It is a comprehensive index of structural strain and damage. , , and For the corresponding weights, and ; S6: Based on the metastable evolution level, output feedback control measures to the vibration table control system, data acquisition system, or test monitoring terminal.

2. The method for metastable state identification and control of cross-fault tunnel lining according to claim 1, characterized in that, The model partitions include the affected section of the hanging wall, the affected section of the footwall, the affected section of the fault fracture zone and / or the affected section of the weak interlayer, and the standard section far from the fault; the lining parts include the arch crown, arch waist, sidewalls and invert arch; the sensors include accelerometers, displacement gauges, earth pressure cells and strain gauges.

3. The method for metastable state identification and control of cross-fault tunnel lining according to claim 1, characterized in that, The comprehensive deformation development index includes the relative displacement deformation index determined by displacement measurement data on both sides of the fault fracture zone and / or weak interlayer; when the relative displacement increment within adjacent loading levels or adjacent time windows exceeds a set threshold, it is determined that there is an accelerated development of displacement deformation in the lining adjacent to the fault.

4. The method for metastable state identification and control of cross-fault tunnel lining according to claim 1, characterized in that, The comprehensive index for earth pressure redistribution includes an index for the development of the difference in earth pressure between the hanging wall and the footwall, determined by earth pressure measurement data from the hanging wall and footwall affected by the fault. When the representative earth pressure difference between the hanging wall and the footwall continues to increase, the ratio of earth pressure on both sides exceeds a set threshold, or the earth pressure on one side decreases while the earth pressure in adjacent areas increases, it is determined that the earth pressure around the lining is developing in a biased manner or that the contact pressure is unloading-transferring.

5. The method for metastable state identification and control of cross-fault tunnel lining according to claim 1, characterized in that, The comprehensive index of dynamic degradation is determined by one or more of the following: change in acceleration amplification factor, reduction in fundamental frequency, or change in damping ratio; the comprehensive index of structural strain damage is determined by one or more of the following: peak strain utilization rate, residual strain, or strain increment.

6. The method for metastable state identification and control of cross-fault tunnel lining according to claim 1, characterized in that, The metastable evolution levels include stable state, damage initiation state, mild metastable state, moderate metastable state, highly metastable state, and near-instability / destruction state.

7. The method for metastable state identification and control of cross-fault tunnel lining according to claim 1, characterized in that, The feedback control measures include: When the damage is in the nascent stage or in a mild metastable state, increase the sampling frequency during critical response periods; When in a moderately metastable state, insert a low-amplitude sweep frequency retest; When in a highly metastable state, reduce the peak ground acceleration of the input seismic wave for the next loading level or suspend loading checks; When the system is in a state close to instability / failure, the loading process is terminated and key response data before and after the failure is saved.

Images