A method and system suitable for diagnosing liquefaction risk of silt layer section of submarine tunnel

Abstract

The present application relates to the field of tunnel safety monitoring and alarm, specifically to a liquefaction risk diagnosis method and system suitable for submarine tunnel silt layer section, aiming at solving the key technical problems of insufficient multi-dimensional feature perception, low multi-source information fusion and lagging early warning response in traditional liquefaction monitoring method. Specifically, it includes: installing a multi-dimensional monitoring network to collect monitoring data; preprocessing the monitoring data to calculate the characteristic value of key physical quantity; the key physical quantity includes the instantaneous change rate of pore water pressure and the correlation coefficient of pore pressure-soil pressure; based on the effective stress principle, sequentially execute joint early warning criterion check; according to the trigger combination of the criterion, execute graded early warning and generate a diagnosis report. This method can accurately identify the liquefaction risk of the tunnel crossing the silt layer section, so as to issue a high-reliability risk diagnosis signal before the liquefaction causes substantial damage to the tunnel structure.

Description

Technical Field

[0001] This invention relates to the field of tunnel safety monitoring and alarm, specifically to a method and system for diagnosing the liquefaction risk in silty sand layers of submarine tunnels. Background Technology

[0002] When a submarine tunnel traverses a silty sand layer, the loose and highly saturated silty sand is prone to liquefaction under external dynamic loads (such as earthquakes, wave impacts, and construction vibrations). Liquefaction causes the soil to lose shear strength, manifested as a sharp increase in excess pore water pressure and a sudden drop in effective stress. This can lead to damage such as tunnel structural uplift, settlement, lateral drift, or joint misalignment, seriously threatening the safe operation of the tunnel.

[0003] Traditional liquefaction monitoring methods often rely on single physical quantities such as settlement or pore water pressure, making it difficult to comprehensively capture the multi-dimensional precursory characteristics of liquefaction evolution. In addition, such methods generally suffer from information silos and delayed warnings, often triggering alarms only when substantial damage such as cracking or misalignment occurs in the tunnel structure, resulting in missed opportunities for optimal intervention.

[0004] To address the aforementioned technical bottlenecks, this invention proposes an early warning system that integrates multi-dimensional monitoring and information fusion technologies. This system not only boasts high reliability and low cost but also exhibits excellent engineering applicability, thereby effectively achieving early and accurate identification and warning of liquefaction risks. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a method and system for diagnosing liquefaction risks in silty sand layers of submarine tunnels. This system effectively overcomes key technical challenges in traditional liquefaction monitoring methods, such as insufficient multi-dimensional feature perception capabilities, low multi-source information fusion, and delayed early warning response.

[0006] To achieve the above objectives, this invention proposes a method for diagnosing liquefaction risk in silty sand layers of submarine tunnels, which specifically includes the following steps:

[0007] (1) Install a multi-dimensional monitoring network to collect monitoring data;

[0008] (2) The monitoring data is preprocessed to calculate the characteristic values ​​of key physical quantities; the key physical quantities include the instantaneous change rate of pore water pressure and the correlation coefficient between pore pressure and earth pressure;

[0009] (3) Based on the effective stress principle, perform joint early warning criterion checks sequentially;

[0010] (4) Based on the trigger combination of the criteria, execute the graded warning and generate a diagnostic report.

[0011] Based on the above method, the present invention also provides a liquefaction risk diagnosis system for silty sand layer sections of submarine tunnels, including: a multi-dimensional monitoring network, a data acquisition and transmission module, and a central processing unit.

[0012] This invention constructs a method and system for diagnosing liquefaction risks in silty sand sections of submarine tunnels, enabling accurate judgment of liquefaction occurrence in tunnel sections traversing silty sand layers, thereby issuing highly reliable risk diagnosis warnings before liquefaction causes substantial damage to the tunnel structure.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0014] This invention achieves full-chain monitoring of liquefaction risk—from "inducing factors (increased pore pressure)" to "mechanical mechanisms (changes in effective stress)" and then to "structural responses (deformation / joint opening)"—by constructing a multi-dimensional monitoring network encompassing pore water pressure, contact soil pressure, and structural response in the silty sand layer section of a submarine tunnel. This multi-parameter collaborative diagnostic mode can comprehensively capture the precursory information and dynamic evolution process of liquefaction, overcoming the limitations of traditional single monitoring methods.

[0015] This invention abandons the traditional method that relies on the absolute value threshold of a single physical quantity, and establishes a hierarchical early warning criterion system based on "drastic changes in pore pressure," "anomaly in the linkage between pore pressure and earth pressure," and "coupling of hydraulic and structural responses." By analyzing the coordinated anomaly patterns of multiple related physical quantities in time and space for cross-validation, false alarms caused by single sensor failures, environmental noise interference, or local anomalies are effectively eliminated, significantly improving the reliability of risk identification.

[0016] This invention designs a multi-source information cross-validation and data reconstruction mechanism. Even if some sensors experience temporary failures or data loss, the system can still maintain the normal operation of the core diagnostic functions using monitoring data from other dimensions, avoiding system paralysis caused by single-point failures and demonstrating strong environmental adaptability and engineering reliability.

[0017] This invention, based on mature sensor technology and efficient data fusion logic, eliminates the need for complex numerical simulations or parameter inversion. The system is simple to construct, cost-effective, and easily applicable to existing tunnel projects. Furthermore, the system can sensitively capture weak signals in the early stages of liquefaction, providing valuable time for operational adjustments and preventative measures. This represents a shift from "passive post-disaster response" to "proactive pre-disaster warning," effectively preventing structural damage and safety accidents, and ensuring the long-term operational safety and service life of the subsea tunnel. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the system workflow of the present invention;

[0019] Figure 2This is a schematic diagram of the built-in judgment logic of the present invention. Detailed Implementation

[0020] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0021] like Figure 1 As shown, a method and system for diagnosing liquefaction risks in silty sand layers of submarine tunnels includes the following specific implementation steps:

[0022] S1: Sensor Arrangement and Installation

[0023] Based on the engineering geological survey report and numerical simulation results, the monitoring range of the tunnel silt layer section is accurately delineated, and a differentiated cross-section layout mechanism based on risk level is established: the basic cross-section spacing in general sections is set at 10~20m, while in high-risk or geologically abrupt sections such as shallow burial and thick silt layers, the spacing is increased to 5~10m.

[0024] For pore water pressure monitoring, corrosion-resistant probe-type sensors are pre-embedded behind the lining at the arch crown, arch waist, and arch bottom of each section, and are embedded into the soil through drilling to ensure direct contact. At the same time, soil pressure and pore pressure are coupled and monitored in the same section, with three soil pressure cells placed at the lining-surrounding rock interface in each section to achieve synchronous sensing of stress-seepage field.

[0025] For structural response monitoring, the system integrates a hydrostatic level (monitoring settlement / heave), a key section inclinometer (monitoring lateral drift), and a segment joint gauge (monitoring opening / misalignment) along the longitudinal direction. To optimize resource allocation, the density of structural monitoring points is dynamically adjusted: in general sections, to ensure timely monitoring, monitoring points are spaced two sections apart, i.e., 20-40m. In high-risk sections, monitoring points are spaced 10-20m apart, thus ensuring timely monitoring while achieving early and accurate warnings of liquefaction risks and structural damage.

[0026] S2: Data Acquisition and Transmission

[0027] This system constructs a dual-mode redundant transmission network combining a wired RS485 bus (main link) and a wireless LoRaWAN (backup link). This network features intelligent fault switching, automatically activating the backup link when the main link fails, effectively mitigating the risk of single-point failure.

[0028] At the data acquisition level, the system introduces a risk-triggered adaptive frequency conversion acquisition mechanism. Under normal conditions, the system operates at a base frequency of 1Hz to achieve a balance between real-time performance and resource consumption; once liquefaction precursors or structural anomalies are detected, it immediately and automatically switches to a 10Hz high-frequency mode to capture critical transient processes at high resolution.

[0029] To address the physical characteristics of each monitored parameter, the system employs differentiated frequency configurations: pore water pressure and soil pressure, due to the rapid stress-seepage coupling changes involved in the liquefaction process, require synchronous high-frequency acquisition at 1Hz–10Hz to support real-time effective stress analysis; while structural displacement and joint deformation, due to response lag, can be captured using frequencies of 0.1Hz–1Hz. This strategy ensures the integrity of key data while significantly improving the overall system efficiency.

[0030] S3: Central Processing Unit and Algorithm Implementation

[0031] The hardware platform of this system is built on an industrial-grade high-performance computer, compatible with Linux and Windows operating system environments. To meet the high real-time requirements of multi-source information fusion algorithms, the computing unit is equipped with a multi-core Intel or AMD processor with a main frequency of ≥3.0GHz, supporting large-scale parallel computing; for memory, ≥16GB of DDR4 2666MHz industrial-grade modules are selected to ensure low latency and high stability of data processing. The storage system adopts a tiered storage strategy, with a 256GB industrial-grade SSD providing high-speed cache for system operation, and supports expansion to 1TB HDD for persistent storage of massive historical data.

[0032] At the software level, the system utilizes mainstream Python or C++ compilation environments to independently develop a multi-source information fusion diagnostic algorithm. This algorithm mainly includes the following modules:

[0033] The data preprocessing module is primarily responsible for cleaning and optimizing the raw sensor data to improve the accuracy and reliability of subsequent diagnostics. To suppress high-frequency noise and power frequency interference while retaining the true low-frequency signals reflecting the liquefaction process, a digital low-pass filter is used for signal filtering; simultaneously, a moving average window method is employed for data smoothing and noise reduction.

[0034] Feature Extraction and Multi-Source Information Fusion Diagnostic Module: This module first performs in-depth mining of multi-source heterogeneous data such as pore water pressure, soil pressure, structural displacement, and joint deformation. It extracts key statistical features such as change rate, evolution trend, and peak / extreme values ​​through algorithms to construct a high-dimensional feature vector. Subsequently, based on the effective stress principle of soil mechanics and the liquefaction evolution mechanism, the module uses a multi-source information fusion algorithm to comprehensively evaluate the feature vector. The system constructs a rigorous three-level progressive joint early warning criterion, capable of dynamically identifying the initiation, development, and critical state of liquefaction risk, thereby achieving highly reliable intelligent risk decision-making.

[0035] Specifically, the feature extraction process focuses on capturing the mechanical response essence of the liquefaction process, and the extracted features include the following three categories:

[0036] ① Transient characteristics (used to capture induced signals)

[0037] Instantaneous rate of change of pore water pressure: Calculated by dividing the difference between the pore water pressure at the current moment and the previous moment by the sampling interval. This feature is a direct input to criterion one, used to identify "step-like jumps" or "pulse-like spikes" in excess static pore water pressure caused by external dynamic loads, which are the earliest inducing signals of the liquefaction process.

[0038] ② Linkage characteristics (used to verify mechanical mechanisms)

[0039] Pearson correlation coefficient between pore pressure and earth pressure: Within a fixed time window (e.g., 60 seconds) after criterion one is triggered, the correlation coefficient between the pore water pressure sequence and the total earth pressure sequence within that window is calculated. This characteristic is the basis for criterion two.

[0040] ③ Response characteristics (used to confirm structural consequences)

[0041] Structural displacement increment: After triggering criterion one or two, a subsequent monitoring time window (e.g., 300 seconds) is opened. Calculate the cumulative increment of the tunnel's vertical displacement, lateral displacement, and joint changes relative to the trigger time within this window.

[0042] Multi-source information fusion diagnosis is to diagnose the features extracted above. Its core is to implement three-level joint early warning criteria, namely, pore pressure drastic change criterion, pore pressure-earth pressure linkage anomaly criterion, and hydraulic-structural response coupling criterion.

[0043] The warning criteria of this system are based on the effective stress principle in soil mechanics. This principle states that the shear strength of saturated soil depends on its effective stress, and its core mathematical expression is:

[0044] ,

[0045] in, For a moment The effective stress of the soil (kPa), For a moment The total stress (kPa) was obtained in real time by earth pressure cell monitoring. For a moment The pore water pressure (kPa) is synchronously detected and obtained by a pore water pressure sensor.

[0046] When liquefaction occurs, the pore water pressure Rise sharply and approach the total stress This leads to effective stress When the shear strength of the soil decreases significantly or even approaches zero, the soil loses its shear strength. This system accurately captures the critical state of liquefaction germination by calculating this difference in real time.

[0047] The system's built-in judgment logic is as follows: After judgment one is triggered, judgment two and judgment three are immediately performed. If neither judgment two nor judgment three is triggered, a level one warning is initiated. If judgment two is triggered but judgment three is not triggered, a level two warning is initiated. If both judgment two and judgment three are triggered, a level three warning is initiated. At the same time, if judgment two is not triggered but judgment three is triggered, a level three warning will also be initiated.

[0048] Criterion 1: Criterion for Dramatic Changes in Pore Pressure

[0049] To identify whether there is a sharp increase in pore water pressure with liquefaction characteristics, the system calculates the instantaneous rate of change of pore water pressure in real time. :

[0050] ,

[0051] in, The data sampling interval, For a moment The pore water pressure (kPa).

[0052] This criterion is used to identify the cumulative effect of excess pore water pressure in silt layers induced by external dynamic loads. Its core logic is: when the instantaneous rate of change of pore water pressure... Breaking the threshold At that time, it was determined to be a precursor to liquefaction risk.

[0053] To ensure the universality and accuracy of the criterion, the threshold... The determination of the parameters employs a data-driven adaptive statistical method. First, the initial range is defined based on data inversion and high-precision numerical simulation of historical typical liquefaction cases. Specifically, pore water pressure data is collected during the system's quiescent period without external dynamic loads, and the standard deviation of its rate of change is calculated. The initial value of the threshold is set to 3. Up to 5 .

[0054] Let the rate of change of pore water pressure be monitored during the quiescent period. Standard deviation The threshold is then set as follows:

[0055] .

[0056] Subsequently, the initial values ​​were iteratively fine-tuned and calibrated based on the on-site in-situ test data, and finally the optimal discrimination criteria adapted to the geological characteristics of this project were established.

[0057] To further reduce the risk of false alarms and reserve management redundancy, this invention innovatively proposes a two-layer threshold linkage mechanism: a attention threshold (lower) and an action threshold (higher). When the attention threshold is exceeded, the system records it as "abnormal," prompting engineers to pay attention, but not necessarily triggering a high-level alarm. A liquefaction warning is only triggered when the action threshold is exceeded and other criteria also show abnormalities. In this invention... This is the action threshold.

[0058] Criterion 2: Criterion for Anomaly in the Linkage between Pore Pressure and Earth Pressure

[0059] To verify whether the increase in pore pressure leads to the loss of effective stress, and to provide mechanical evidence for liquefaction, at the precise moment when criterion one is triggered (denoted as...). The system will immediately perform the following operations:

[0060] 1. with Centered on the data, a continuous segment of historical monitoring data is extracted in both forward and backward directions;

[0061] 2. Extract a time window Tw of a pre-defined fixed length (e.g., Tw = 60 seconds);

[0062] 3. Obtain two data sequences: in the window [ −Tw / 2, Within [+Tw / 2], extract the pore water pressure sequence P and the total earth pressure sequence S, respectively. Calculate their Pearson correlation coefficient. :

[0063] ,

[0064] in, For covariance, and The standard deviation is denoted as .

[0065] when When a strong negative correlation is observed, criterion two is triggered.

[0066] From the perspective of soil mechanics, this strong negative correlation is a hallmark characteristic of liquefaction: it represents the excess pore water pressure under dynamic loading. The accumulation and rise of soil effective stress The attenuation decreases exhibit a highly synchronized reverse evolution trend. This synchronicity directly leads to the effective stress... A sudden drop below the critical value within a short period of time indicates that the soil is about to lose its shear strength.

[0067] Criterion 3: Hydraulic-Structural Response Coupling Criterion

[0068] To confirm that liquefaction has had observable mechanical effects on the tunnel structure, the triggering time of criterion one was used. Starting from this point, open the monitoring time window. Within this window, criterion three is triggered if the structural displacement satisfies any of the following conditions:

[0069] Vertical displacement satisfy: ;

[0070] Lateral displacement satisfy: ;

[0071] seam variable satisfy: ;

[0072] Regarding the determination of the vertical displacement threshold, this invention is based on the minimum radius of the tunnel's longitudinal deformation curve. The geometric mapping relationship between them is derived, and a safety factor is introduced for correction, thereby determining the final warning threshold.

[0073] Specifically, for a length of For tunnel sections with a distance between monitoring points (i.e., the allowable differential settlement threshold), the allowable differential settlement threshold is... It can be estimated using the following formula:

[0074] ,

[0075] in, The vertical displacement threshold defined in this invention ; The spacing between monitoring points (m) is a known parameter determined by the deployment scheme of this monitoring system. The minimum radius of curvature (m) of the longitudinal deformation curve of the tunnel is a safety control parameter that must be met as specified in the tunnel design documents.

[0076] The lateral displacement threshold can be determined by combining the following two methods:

[0077] a) Back calculation based on the allowable deviation of the design axis:

[0078] Obtain the maximum allowable deviation of the tunnel's axis and horizontal position from the tunnel design documents. This value represents the ultimate limit for ensuring tunnel alignment and safety. Considering the preventative nature of liquefaction warnings, the lateral displacement threshold should be set to 1 / 3 to 1 / 5 of this ultimate deviation, i.e.:

[0079] .

[0080] b) Based on the empirical proportional relationship with the vertical displacement threshold:

[0081] In geotechnical engineering practice, because soil is not an isotropic medium and its restoring force differs from that in the vertical direction, the horizontal deformation tolerance is generally considered lower than that in the vertical direction. Therefore, based on reliable empirical relationships verified by extensive engineering practice, the lateral displacement threshold is usually set to 1 / 2 to 2 / 3 of the vertical displacement threshold. Specifically:

[0082] ,

[0083] Lateral displacement threshold in this invention The final value was determined by cross-validating the two methods (a) and (b) mentioned above, and the smaller value was taken as the final adopted value. This ensures that the system has extremely high sensitivity to the major risk of tunnel lateral instability.

[0084] Threshold for variation of segment joints Its determination is directly based on the calculation in the "Specifications for Design of Underwater Tunnels on Highways", as shown in the formula in the specifications:

[0085] ,

[0086] in, The maximum opening (m) of the inter-ring joint between the segments. Let be the minimum radius (m) of the tunnel's longitudinal deformation curve. The outer diameter of the tunnel (m) The width of the tunnel segment ring (m) The circumferential gap (m) caused by generation and construction errors. This refers to the joint opening (m) caused by later deformation of the tunnel.

[0087] Considering that liquefaction warnings aim to prevent structures from reaching their limit states, this invention needs to set a warning threshold. Set to be much smaller than the theoretical maximum allowable value . Specifically, take One-third to one-half of the value is used as the warning threshold. For example, if the calculated maximum allowable opening amount... Then the threshold for seam variation It should be set to 2-3mm. This ensures that the system can issue an alarm before the joint opening becomes too large but has not yet compromised the structural waterproofing.

[0088] Algorithm Optimization Module: This module aims to construct a data-driven dynamic threshold evolution system. Through real-time statistical learning and feedback control, it achieves adaptive adjustment of diagnostic sensitivity and continuous optimization of the system's false alarm / false negative rate. Specifically, it includes the following two aspects:

[0089] 1. Statistical Adaptive Mechanism Based on Rolling Time Window

[0090] The system provides key diagnostic features (such as the rate of change of pore water pressure). Maintain a fixed-length first-in-first-out (FIFO) data buffer to store data from a past period. (e.g., 24 hours) All characteristic values ​​within a no-alarm state. The module periodically calculates the moving average of the data in the buffer. and moving standard deviation And based on this, the threshold in criterion one is dynamically updated, for example:

[0091] ,

[0092] in, This is the sensitivity coefficient.

[0093] In the initial stage of system deployment, adaptive initialization is performed based on the kurtosis characteristics of historical data from the stable operation phase. Value. If the data distribution exhibits heavy-tailed characteristics (high kurtosis, indicating frequent extreme fluctuations), then set a larger initial value. Set a value (e.g., K=5) to enhance noise robustness; if the distribution exhibits light-tailed characteristics (low kurtosis, concentrated distribution), set a smaller initial value. Values ​​(e.g., K=3.5) can be used to improve the sensitivity of capturing weak signals.

[0094] The system incorporates a dual-loop feedback mechanism for false alarms and false negatives. By periodically analyzing false alarm logs, if the false alarm rate consistently exceeds a preset upper limit, the algorithm automatically increases the K value to suppress false alarms. Simultaneously, by performing backtracking analysis based on confirmed real liquefaction events, if it is found that a risk is missed due to an excessively large K value, the K value is automatically lowered, thereby achieving a dynamic balance between the false alarm rate and the false negative rate.

[0095] 2. Multi-source criterion linkage optimization strategy based on risk conservation

[0096] To optimize the structural response threshold, this invention proposes a linked indirect optimization strategy. This strategy is achieved by introducing a dynamic synergistic factor for each structural response threshold; this factor does not directly correct the theoretical benchmark threshold, but is embedded as a weighting factor in the judgment logic of criterion three.

[0097] Taking vertical displacement as an example, the optimized criterion three trigger condition becomes:

[0098] ,

[0099] Synergistic Factor The optimization of the sensitivity coefficient K for pore pressure change rate is synchronously and negatively correlated. The specific control strategy is as follows:

[0100] 1. "Cautious Premonitions, Sensitive Consequences" Mode: When false alarm log analysis indicates the need to increase the K value (i.e., to strengthen the stringency of the pore pressure precursor criterion), the system synchronously and appropriately reduces it. Value. Due to As a multiplicative weight, its reduction means lowering the trigger threshold for structural response, making criterion three more sensitive. This strategy aims to compensate for the front-end filtering effect: when the detection of early warning signs becomes more cautious, it is necessary to enhance the sensitivity to the "consequences" of structural damage to prevent the underreporting of real risks due to excessively high front-end thresholds.

[0101] 2. "Sensitive Precursors, Forgiving Consequences" Mode: When retrospective analysis confirms a real event and decides to reduce the K value (i.e., increase the sensitivity of the precursor criterion), the system synchronously and appropriately increases the K value. This raises the trigger threshold for the structural response, making criterion three relatively "tolerant." This strategy aims to suppress combined false alarms: avoiding new combinations of false alarms arising from both front-end and back-end criteria being in a highly sensitive state simultaneously.

[0102] To prevent excessive parameter drift from causing system instability, all cooperative factors are... By setting operational boundaries, this constraint mechanism ensures that the structural response threshold is always dynamically adjusted within the designed safety margin, thus guaranteeing the engineering reliability of the system.

[0103] Early warning decision and report generation module: Based on the combination of triggered criteria, execute the hierarchical early warning logic and generate a report.

[0104] When the system issues an alarm, maintenance personnel receive a notification through the monitoring platform.

[0105] Level 1 alert is a risk warning: when only criterion 1 is triggered, the system marks the segment as a "pore pressure response sensitive area" to prompt maintenance personnel to pay attention, but no immediate action is required.

[0106] Level 2 alarm is a high-risk alarm: When criterion 1 and criterion 2 are triggered simultaneously, the system issues a "high-risk liquefaction alarm", suggesting that monitoring be strengthened and emergency measures be prepared.

[0107] Level 3 alarm is a confirmation alarm: when any criterion and criterion 3 are triggered simultaneously, the system issues a confirmation alarm that "liquefaction may have occurred and the structure is being affected," and recommends immediate manual inspection and intervention, such as adjusting the operation plan and taking targeted reinforcement measures.

[0108] The system automatically generates diagnostic reports, clearly indicating the sensor data, time, location, and corresponding physical phenomena that triggered the alarm, supporting decision analysis.

[0109] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for diagnosing liquefaction risk in silty sand layers of submarine tunnels, characterized in that, Includes the following steps: S1: Install a multi-dimensional monitoring network to collect monitoring data; S2: Preprocess the monitoring data and calculate the characteristic values ​​of key physical quantities; The key physical quantities include the instantaneous rate of change of pore water pressure and the correlation coefficient between pore pressure and earth pressure. S3: Based on the effective stress principle, perform joint early warning criterion checks sequentially; S4: Based on the trigger combination of the criteria in step S3, execute the hierarchical warning and generate a diagnostic report.

2. The method for diagnosing liquefaction risk in silty sand layers of a submarine tunnel according to claim 1, characterized in that, The joint early warning criterion check in S3 includes: Criterion 1: Triggered when the instantaneous rate of change of the pore water pressure exceeds its dynamic action threshold; Criterion 2: Near the time point triggered by Criterion 1, calculate the Pearson correlation coefficient between the pore water pressure sequence and the total earth pressure sequence. Triggering occurs when the coefficient is less than -0.

7. Criterion 3: Starting from the trigger time of Criterion 1 or Criterion 2, if the increase in vertical displacement, lateral displacement or joint change exceeds their respective dynamic thresholds within the subsequent monitoring time window, it will be triggered.

3. The method for diagnosing liquefaction risk in silty sand layers of a submarine tunnel according to claim 2, characterized in that, The determination of the initial value of the action threshold in the first criterion includes: collecting pore water pressure data of the system during a quiescent period without external dynamic load, and calculating the standard deviation of its rate of change. The initial value of the action threshold is set to 3. Up to 5 .

4. The method for diagnosing liquefaction risk in silty sand layers of a submarine tunnel according to claim 2, characterized in that, The vertical displacement threshold in the third criterion The method for determining it includes the following steps: (1) Obtain the minimum radius of curvature of the longitudinal deformation curve of the tunnel. and the distance between monitoring points ; (2) Based on the minimum radius of curvature and the distance between monitoring points Using the formula The differential settlement threshold was calculated. ; (3) The differential settlement threshold A safety factor correction is performed, and the corrected value is determined as the vertical displacement threshold.

5. The method for diagnosing liquefaction risk in silty sand layers of a submarine tunnel according to claim 2, characterized in that, The third criterion includes the lateral displacement threshold. The value is the smaller of the following two calculation results: (1) Maximum allowable deviation of tunnel design axis 1 / 3 to 1 / 5; (2) 1 / 2 to 2 / 3 of the vertical displacement threshold.

6. The method for diagnosing liquefaction risk in silty sand layers of a submarine tunnel according to claim 2, characterized in that, The method for determining the dynamic action threshold in criterion one includes: dynamically calculating the moving average value based on the characteristic value data of the no-alarm state within the rolling time window. and moving standard deviation And based on this, according to the formula The action threshold in criterion one is dynamically updated, where K is a sensitivity coefficient that is adaptively adjusted based on false alarm logs and backtracking analysis of real events.

7. The method for diagnosing liquefaction risk in silty sand layers of a submarine tunnel according to claim 2, characterized in that, The method for determining the dynamic threshold in criterion three includes: (1) Introduce synergistic factors for the vertical displacement threshold, the lateral displacement threshold, and the joint variation threshold, respectively. ; (2) When adjusting the sensitivity coefficient Simultaneously and in reverse, the corresponding synergistic factors are adjusted. This allows for a risk balance between the sensitivity of criterion one and criterion three.

8. A system applicable to the liquefaction risk diagnosis method for silty sand layers in submarine tunnels as described in any one of claims 1-7, characterized in that, include: Multi-dimensional monitoring network: including pore water pressure sensor, earth pressure cell, hydrostatic level, inclinometer and joint gauge, used to collect real-time data on pore water pressure, total earth pressure, vertical displacement, horizontal displacement and segment joint opening and closing degree. Data acquisition and transmission module: connected to the multi-dimensional monitoring network, configured to acquire data from the monitoring network at a predetermined frequency and transmit it to the central processing unit; Central processing unit: Communicatively connected to the data acquisition and transmission module, with built-in memory and processor. The processor is configured to run the following functional modules to execute the liquefaction risk diagnosis process: data preprocessing module, feature extraction and multi-source information fusion diagnosis module, algorithm optimization module, and early warning decision and report generation module.

9. A system for liquefaction risk diagnosis in silty sand layers of submarine tunnels according to claim 8, characterized in that, The processor includes the following functional modules: The data preprocessing module is configured to filter and denoise the received time-series data and remove outliers, and output a standardized monitoring data sequence. The feature extraction and multi-source information fusion diagnostic module is configured to calculate the instantaneous change rate of pore water pressure and the correlation coefficient between pore pressure and earth pressure from standardized data, and to execute joint early warning criteria. The algorithm optimization module is configured to perform dynamic evolution and joint optimization of the execution threshold; The early warning decision and report generation module is configured to execute tiered early warnings and generate reports based on the results of triggering judgments.

Images