A multimodal bridge vibration monitoring system

Abstract

The present invention relates to the technical field of bridge monitoring, and provides a multi-modal bridge vibration monitoring system; the system includes a data acquisition and processing module, a bridge monitoring module, an optimized monitoring accuracy module, and a bridge health assessment and early warning module; through the bridge safety integration network, intelligent analysis and health status assessment of multi-dimensional data are realized; the system uses the k,r dual-parameter optimization algorithm to dynamically adjust the hyperparameters of the bridge safety integration network, optimize the learning rate, regularization parameter, and batch size to adapt to different bridge environments and operating conditions, and improve the monitoring accuracy and long-term stability; through multi-modal data fusion, intelligent optimization, and adaptive adjustment, the present invention improves the comprehensiveness, accuracy, and adaptability of bridge vibration monitoring, and provides an efficient and scientific technical guarantee for bridge safety management.

Description

Technical Field

[0001] The present invention relates to the technical field of bridge monitoring, and in particular to a multimodal bridge vibration monitoring system. Background Art

[0002] With the rapid development of modern transportation infrastructure, bridges, as an important part of highway and railway transportation, their structural safety and service life have a significant impact on public safety and social economy. However, the traditional bridge vibration monitoring system has the following deficiencies: First, the traditional bridge vibration monitoring system mainly relies on a single type of sensor, making it difficult to comprehensively reflect the health status of the bridge, resulting in one-sided monitoring data and affecting the accuracy of damage identification; Second, the traditional system fails to make full use of the multi-source information collected by sensors and lacks an effective data fusion strategy. The observation results of different sensors are difficult to be processed collaboratively, leading to data redundancy and information loss, and affecting the reliability of monitoring results; Finally, the traditional system is difficult to effectively capture the spatio-temporal correlation characteristics of bridge vibration data; especially when the bridge suffers from early damage, the traditional methods are often difficult to accurately identify and give early warnings, resulting in a lag in maintenance decisions, increasing the bridge safety risk and maintenance cost. Summary of the Invention

[0003] The present invention proposes a multimodal bridge vibration monitoring system, aiming to improve the accuracy and real-time performance of bridge vibration monitoring, and ensure the structural safety and long-term stable operation of the bridge; the system collects multimodal vibration data of the bridge in real time by arranging accelerometers, strain gauges, displacement sensors, temperature sensors and anemometers on the bridge structure, and constructs a bridge safety fusion network by using topology perception, multi-layer gated fusion and spectrum feature extraction to capture the health status of the bridge from multiple dimensions; the system first establishes a sensor interaction graph based on the spatial dependence, time dependence and modal dependence of the bridge sensor data, and initializes the hyperparameters of the bridge safety fusion network based on the topological structure to make it adapt to the vibration characteristics under different bridge environments; in the data analysis process, the system uses the bridge safety fusion network to process the multimodal data, extracts low-frequency and high-frequency features from the frequency domain, and accurately identifies the long-term structural changes and sudden damages of the bridge; to further optimize the monitoring accuracy, the system introduces a k,r double-parameter optimization algorithm to dynamically adjust the hyperparameters of the bridge safety fusion network, optimize the learning rate, momentum coefficient, regularization parameter and batch size, and ensure that the monitoring system can still maintain high-efficiency and accurate bridge health status prediction and vibration monitoring capabilities under complex loads and environmental interferences; the present invention constructs a set of efficient, intelligent and stable bridge vibration monitoring system through multimodal data fusion, intelligent feature extraction and adaptive optimization strategies, provides scientific support for bridge structural health management, and improves the bridge safety guarantee level.

[0004] The present invention provides a multimodal bridge vibration monitoring system, which includes a bridge sensor, a data acquisition module, a bridge monitoring module, an optimized monitoring accuracy module, and a bridge health assessment and early warning module;

[0005] The data acquisition module arranges bridge sensors on the bridge. The bridge sensors include accelerometers, strain gauges, displacement sensors, temperature sensors, and anemometers; it collects acceleration data, displacement data, strain data, temperature data, and wind speed data through the bridge sensors to obtain multimodal bridge data;

[0006] The bridge monitoring module constructs a bridge safety fusion network through topology perception, multi-layer gated fusion, and spectral feature extraction, and initializes the hyperparameters of the bridge safety fusion network; it processes the multimodal bridge data through the bridge safety fusion network for bridge vibration monitoring and generates a prediction result of the bridge health status;

[0007] The optimized monitoring accuracy module constructs a k,r two-parameter optimization algorithm through Bregman divergence-guided adaptive gradient optimization, k,r-exponential transformation, and a generalized regularization mechanism, and optimizes the hyperparameters of the bridge safety fusion network through the k,r two-parameter optimization algorithm;

[0008] The bridge health assessment and early warning module: generates a bridge health status report based on the prediction result of the bridge health status, including the degree of structural damage, the decay rate of bearing capacity, the change rate of vibration mode, the risk of displacement exceeding the limit, the cumulative degree of fatigue damage, the structural stability index, the health status score, and the early warning level.

[0009] Furthermore, the process of the bridge monitoring module generating the prediction result of the bridge health status specifically includes the following steps:

[0010] Step S1: Establish the topological structure of the multimodal bridge data, capture the spatial dependence, time dependence, and modal dependence of the bridge sensors, and construct a sensor interaction graph;

[0011] Spatial dependence: The adjacent sensors on the bridge are interconnected; the sensors include accelerometers, strain gauges, displacement sensors, temperature sensors, and anemometers;

[0012] Time dependence: Establish a connection between the multimodal bridge data at the current moment and the multimodal bridge data in the past;

[0013] Method: Capture the correlation between the multimodal bridge data at the current moment and the multimodal bridge data in the past through an LSTM time series model;

[0014] Association: The multi-modal bridge data at the current moment and the multi-modal bridge data at past moments are used as the input of the LSTM time series model; by learning the relationship between the data at the current moment and the historical data, the LSTM time series model automatically models the dependence of multi-modal bridge data in the time dimension, so as to understand how the data at different time points influence and relate to each other;

[0015] Modal dependence: Different multi-modal bridge data are interrelated;

[0016] Method: The attention mechanism is used to capture the modal dependence between different multi-modal bridge data; the attention mechanism can assign different importance weights to different modal data, so as to better capture the relationship between modalities;

[0017] Association: The attention mechanism processes different multi-modal bridge data in parallel, independently processes the data of each modality, and then fuses them to learn the association between them; in this way, the attention mechanism can identify the relationship between acceleration data, displacement data, strain data, temperature data and wind speed data, and further reveal the mutual influence between modal data;

[0018] Step S2: Calculate the Laplacian matrix using the sensor interaction graph, perform eigen decomposition to obtain low-frequency features and high-frequency features; construct a multi-layer adaptive gating fusion mechanism through a three-layer gating mechanism, dynamic attention weighting and gradient constraint; use the multi-layer adaptive gating fusion mechanism to perform feature fusion on the low-frequency features and high-frequency features to generate spectral domain fusion features;

[0019] Step S3: Recover the missing data of the spectral domain fusion features, and use the multi-head attention mechanism to optimize the feature fusion to generate complete bridge data;

[0020] Step S4: Combine the modal recovery loss function and the cross-entropy loss function to construct a bridge health classifier; input the complete bridge data into the bridge health classifier to generate the prediction result of the bridge health state.

[0021] Furthermore, optimize the monitoring accuracy module and the process of optimizing the hyperparameters of the bridge safety fusion network, which specifically includes the following steps:

[0022] Step B1: Define the search space of the hyperparameters of the bridge safety fusion network, and the search space includes learning rate, momentum coefficient, regularization parameter and batch size; set the k,r double parameters to optimize the hyperparameters of the bridge safety fusion network; k controls the smoothness of the hyperparameter change, and r controls the sensitivity of the hyperparameter adjustment;

[0023] Step B2: Define the loss function and calculate the divergence gradient of the hyperparameters of the bridge safety fusion network with respect to the loss function;

[0024] Step B3: Combine the divergence gradient and use the k, r dual parameters to update the hyperparameters of the bridge safety fusion network;

[0025] Step B4: Set the convergence condition. If the convergence condition is met, stop the optimization; otherwise, return to Step B2 to continue updating the hyperparameters of the bridge safety fusion network.

[0026] Further, Step B3 specifically includes the following steps:

[0027] Step B31: Optimize the learning rate and momentum coefficient using the k, r-exponential gradient;

[0028] Step B32: Optimize the regularization parameter using the k, r-generalized regularization;

[0029] Step B33: Update the batch size using the k, r-additive hyperparameter;

[0030] Step B34: Use the Bregman divergence to provide a global optimization strategy, collaboratively optimize the learning rate, momentum coefficient, regularization parameter, and batch size, optimize the high-dimensional hyperparameter search, and ensure the best adaptation of the hyperparameter combination.

[0031] Adopting the above solution, the beneficial effects obtained by the present invention are as follows:

[0032] The present invention realizes the efficient acquisition and intelligent fusion of multi-modal vibration data of bridges, and improves the comprehensive perception ability of the health status of bridge structures; through topology perception, multi-layer gating fusion, and spectrum feature extraction, the system can make full use of multi-modal data such as acceleration, strain, displacement, temperature, and wind speed to construct a bridge safety fusion network, enabling bridge vibration monitoring not to be limited to a single data source, but to conduct comprehensive analysis based on multi-dimensional information; this data fusion method effectively solves the problems of information isolation and one-sided monitoring data in traditional systems, improves the system's ability to analyze the dynamic characteristics of bridges, and makes the monitoring results more accurate and reliable.

[0033] The present invention optimizes the adaptive ability of the bridge vibration monitoring system and enhances the monitoring stability in complex environments; for the structural characteristics and environmental conditions of different bridges, the system uses the k, r dual-parameter optimization algorithm to dynamically adjust the hyperparameters of the bridge safety fusion network, realizing the adaptive optimization of the learning rate, regularization parameter, and batch size; this optimization mechanism ensures that the system can still accurately extract vibration modes and timely detect damage changes of bridges in the face of different loads, temperature changes, and wind speed effects; compared with traditional systems with fixed hyperparameters, this system can dynamically adjust the parameter configuration to adapt to various bridge conditions, significantly improving the long-term stability and adaptability of monitoring.

[0034] In summary, through the combination of multi-modal data fusion and intelligent optimization algorithms, the present invention constructs a set of efficient, accurate, and stable bridge vibration monitoring system; compared with traditional systems, the present invention not only improves the comprehensiveness and accuracy of bridge structural health monitoring, but also optimizes the long-term stability and adaptability of the monitoring system, providing more scientific and efficient technical support for bridge operation and maintenance management. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a schematic diagram of the modules of a multi-modal bridge vibration monitoring system proposed by the present invention;

[0036] Figure 2 It is a schematic structural diagram of the bridge sensor arrangement of the bridge proposed by the present invention;

[0037] Figure 3 It is a schematic flow diagram of step B3 proposed in Embodiment 6. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0038] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0039] Embodiment 1, according to Figure 1 、 Figure 2 The present invention provides a multi-modal bridge vibration monitoring system, which includes a data acquisition module, a bridge sensor, a bridge monitoring module, an optimized monitoring accuracy module, and a bridge health assessment and early warning module;

[0040] The data acquisition module arranges bridge sensors on the bridge. The bridge sensors include accelerometers, strain gauges, displacement sensors, temperature sensors, and anemometers; multi-modal bridge data is obtained by collecting acceleration data, displacement data, strain data, temperature data, and wind speed data through the bridge sensors.

[0041] ;

[0042] The bridge monitoring module constructs a bridge safety fusion network through topology perception, multi-layer gated fusion, and spectral feature extraction, and initializes the hyperparameters of the bridge safety fusion network; the bridge vibration is monitored by processing the multi-modal bridge data through the bridge safety fusion network, and a prediction result of the bridge health status is generated.

[0043] Optimized monitoring accuracy module, constructs a k,r two-parameter optimization algorithm through Bregman divergence-guided adaptive gradient optimization, k,r-exponential transformation and generalized regularization mechanism, and optimizes the hyperparameters of the bridge safety and health integration network through the k,r two-parameter optimization algorithm;

[0044] Bridge health assessment and early warning module: According to the prediction results of the bridge health status, generate a bridge health status report, including the degree of structural damage, the decay rate of bearing capacity, the change rate of vibration mode, the risk of displacement exceeding the limit, the cumulative degree of fatigue damage, the structural stability index, the health status score and the early warning level;

[0045] ;

[0046] Example 2, this example is based on Example 1. In this example, the process of the bridge monitoring module generating the prediction results of the bridge health status specifically includes the following steps:

[0047] Step S1: Establish the topological structure of multi-modal bridge data, capture the spatial dependence, time dependence and modal dependence of bridge sensors, and construct a sensor interaction graph;

[0048] Spatial dependence: Adjacent sensors on the bridge sensors are interconnected; the sensors include accelerometers, strain gauges, displacement sensors, temperature sensors and anemometers;

[0049] Time dependence: Establish a connection between the multi-modal bridge data at the current moment and the multi-modal bridge data in the past;

[0050] Method: Capture the correlation between the multi-modal bridge data at the current moment and the multi-modal bridge data in the past through the LSTM time series model;

[0051] Correlation: Use the multi-modal bridge data at the current moment and the multi-modal bridge data at the past moment as the input of the LSTM time series model; the LSTM time series model automatically models the dependence of multi-modal bridge data in the time dimension by learning the relationship between the data at the current moment and the historical data, so as to understand how the data at different time points affect and correlate with each other;

[0052] Modal dependence: Different multi-modal bridge data are correlated with each other;

[0053] Method: Capture the modal dependence between different multi-modal bridge data through the attention mechanism; the attention mechanism can assign different importance weights to different modal data, so as to better capture the relationship between modalities;

[0054] Association: The attention mechanism processes different multi-modal bridge data in parallel, independently processes the data of each modality, and then fuses them to learn the associations between them. In this way, the attention mechanism can identify the relationships between acceleration data, displacement data, strain data, temperature data, and wind speed data, and further reveal the mutual influences between modal data.

[0055] Step S2: Calculate the Laplacian matrix using the sensor interaction graph, perform eigen-decomposition to obtain low-frequency features and high-frequency features. Construct a multi-layer adaptive gating fusion mechanism through a three-layer gating mechanism, dynamic attention weighting, and gradient constraint. Use the multi-layer adaptive gating fusion mechanism to perform feature fusion on the low-frequency features and high-frequency features to generate spectral domain fusion features. The formula used is as follows:

[0056] ;

[0057] Where, represents the bridge sensor index, represents the adjacent sensor index of the bridge sensor, represents the spectral domain fusion feature, represents the adjacent sensor 's contribution to the bridge sensor ; represents the relationship set, including spatial dependence, temporal dependence, and modal dependence, represents an element in represents the bridge sensor in the relationship 's neighbor set; represents the adaptive weight, controlling the influence of each sensor in the bridge sensor, represents the low-frequency feature of the self-sensor, represents the neighbor node 's low-frequency feature, represents the neighbor node 's high-frequency feature, represents the low-frequency gating, represents the high-frequency gating, represents the modal enhancement gating, represents the feature gradient, represents the feature gradient weight;

[0058] Low-frequency features: Represent the global health state of the bridge, including temperature effects and long-term structural changes;

[0059] High-frequency features: Represent sudden anomalies, including cracks, local damage, and impact vibrations;

[0060] Step S3: Recover the missing data of the spectral domain fusion features, and optimize the feature fusion using the multi-head attention mechanism to generate complete bridge data;

[0061] Step S4: Combine the modal recovery loss function and the cross-entropy loss function to construct a bridge health classifier; input the complete bridge data into the bridge health classifier to generate the bridge health status prediction result. The formula used is as follows:

[0062] Loss function formula:

[0063] ;

[0064] where, represents the modal recovery loss, represents different modes, represents the acceleration data, represents the strain data, represents the displacement data, represents the temperature data, represents the wind speed, represents the total number of samples, represents the sample index, represents the modal dimension, represents the real data, represents the recovered data;

[0065] Classifier formula:

[0066] ;

[0067] where, represents the bridge health status prediction result, represents the cross-entropy loss, represents the bridge health status category, represents the classification weight matrix, represents the bias vector; represents the activation function.

[0068] Embodiment 3. This embodiment is based on Embodiment 1. In this embodiment, the process of the bridge monitoring module generating the bridge health status prediction result specifically includes the following steps:

[0069] Step E1: Establish the topological structure of the multi-modal bridge data, capture the spatial dependence, temporal dependence, and modal dependence of the bridge sensors, and construct a sensor interaction graph;

[0070] Step E2: Calculate the Laplacian matrix using the sensor interaction graph, perform eigenvalue decomposition to obtain low-frequency features and high-frequency features; construct an adaptive fusion model using methods of gating mechanism, attention weighting, and gradient constraint; use the adaptive fusion model to perform feature fusion on the low-frequency features and high-frequency features to generate frequency-domain fusion features;

[0071] Step E3: Recover the missing data in the frequency-domain fusion features, and optimize the feature fusion using the multi-head attention mechanism to generate complete bridge data;

[0072] Step E4: Combine the modal recovery loss function and the cross-entropy loss function to construct a bridge health classifier; input the complete bridge data into the bridge health classifier to generate the prediction result of the bridge health state.

[0073] Example 4: This example is based on Example 2. In this example, the process of optimizing the monitoring accuracy module and optimizing the hyperparameters of the bridge safety fusion network specifically includes the following steps:

[0074] Step B1: Define the search space of the hyperparameters of the bridge safety fusion network. The search space includes the learning rate, momentum coefficient, regularization parameter, and batch size; set the k and r parameters to optimize the hyperparameters of the bridge safety fusion network; k controls the smoothness of the hyperparameter change, and r controls the sensitivity of the hyperparameter adjustment;

[0075] Step B2: Define the loss function and calculate the divergence gradient of the hyperparameters of the bridge safety fusion network with respect to the loss function;

[0076] Step B3: Combine the divergence gradient and use the k and r parameters to optimize and update the hyperparameters of the bridge safety fusion network;

[0077] Step B4: Set the convergence condition. If the convergence condition is met, stop the optimization; otherwise, return to Step B2 to continue updating the hyperparameters of the bridge safety fusion network.

[0078] Example 5: This example is based on Example 2. In this example, the process of optimizing the monitoring accuracy module and optimizing the hyperparameters of the bridge safety fusion network specifically includes the following steps:

[0079] Step U1: Define the search space of the hyperparameters of the bridge safety fusion network. The search space includes the learning rate, momentum coefficient, regularization parameter, and batch size;

[0080] Step U2: Define the loss function and calculate the model loss of the hyperparameters of the bridge safety fusion network;

[0081] Step U3: Update the hyperparameters of the bridge safety fusion network according to the model loss;

[0082] Step U4: Set the convergence condition. If the convergence condition is met, stop the optimization; otherwise, return to Step B2 to continue updating the hyperparameters of the Qiao'an fusion network.

[0083] Example 6. According to Figure 3 , this example is based on Example 4. In this example, Step B3 specifically includes the following steps:

[0084] Step B31: Adaptive gradient update: Use k,r-exponential gradient to optimize the learning rate and momentum coefficient. The formula used is as follows:

[0085] ;

[0086] ;

[0087] Where represents the iteration index, represents the learning rate after the -th round of iterative update, represents the learning rate after the -th round of iterative update; represents the momentum parameter after the -th round of iterative update, represents the momentum parameter after the -th round of iterative update; represents k,r multiplication, used to update the hyperparameters and ; represents the update coefficient, used to control the step size of the learning rate update, represents the update coefficient, used to control the update amplitude of the momentum coefficient, represents the learning rate for the loss function gradient, represents the momentum coefficient for the loss function gradient; represents the exponential function;

[0088] Step B32: Optimize the regularization parameter: Use k,r-generalized regularization to optimize the regularization parameter. The formula used is as follows:

[0089] ;

[0090] Where represents the regularization parameter after the -th round of iteration, represents the current -th round of iteration's regularization parameter, represents k,r-generalized regularization multiplication, ensuring the non-linear update of the regularization parameter; Denote the regularization parameter learning rate, Denote the loss function for gradient, Denote the update factor of the regularization parameter;

[0091] Step B33: Optimize the batch size: Update the batch size using the k,r-additive hyperparameter, and the formula used is as follows:

[0092] ;

[0093] where Denote the batch size after the round of iteration, Denote the current batch size of the round of iteration, Denote the k,r-additive hyperparameter update operation to ensure a smooth adjustment of the batch size; Batch size update step size, Denote the loss function for gradient; Denote the batch size update factor;

[0094] Step B34: Collaborative optimization: Use the Bregman divergence to provide a global optimization strategy, collaboratively optimize the learning rate, momentum coefficient, regularization parameter and batch size, optimize the high-dimensional hyperparameter search, and ensure the best adaptation of the hyperparameter combination.

[0095] Example 7, this example is based on Example 6. In this example, the bridge health assessment and early warning module: Generate a bridge health status report according to the bridge health status prediction result, including the degree of structural damage, the attenuation rate of bearing capacity, the change rate of vibration mode, the risk of displacement exceeding the limit, the cumulative degree of fatigue damage, the structural stability index, the health status score and the early warning level;

[0096] Generate the following bridge health status report:

[0097] ;

[0098] The health status score is "Grade B", the bridge is in a sub-healthy state, and the following measures are taken:

[0099] Inject glue into the crack for crack repair at the crack location;

[0100] Strengthen the fatigue monitoring in the next quarter;

[0101] Maintain the current monitoring frequency and conduct encrypted sampling when the wind load is large;

[0102] If the continuous decrease in modal frequency continues to expand, a static load test shall be carried out to verify the bearing capacity.

[0103] The above describes the present invention and its implementation manners. Such description is not restrictive. What is shown in the drawings is only one of the implementation manners of the present invention, and the actual structure is not limited thereto; generally speaking, if those of ordinary skill in the art are inspired by it and design, without creative efforts, structural manners and embodiments similar to the technical solution without departing from the gist of the present invention, they shall fall within the protection scope of the present invention.

Claims

1. A multi-modal bridge vibration monitoring system, characterized in that: The system includes bridge sensors, data acquisition modules, bridge monitoring modules, optimized monitoring accuracy modules and bridge health assessment and early warning modules; A data acquisition module is used to arrange bridge sensors on the bridge. The bridge sensors include accelerometers, strain gauges, displacement sensors, temperature sensors and anemometers; The acceleration data, displacement data, strain data, temperature data and wind speed data are collected by bridge sensors to obtain multi-modal bridge data; The bridge monitoring module builds a bridge fusion network through topology perception, multi-layer gated fusion and spectrum feature extraction, and initializes the hyperparameters of the bridge fusion network; it processes multi-modal bridge data through the bridge fusion network to monitor bridge vibration and generate bridge health status prediction results; Optimize the monitoring accuracy module, build a k,r dual-parameter optimization algorithm through Bregman divergence-guided adaptive gradient optimization, k,r-exponential transformation and generalized regularization mechanism, and optimize the hyperparameters of the Qiaoan fusion network through the k,r dual-parameter optimization algorithm; Bridge health assessment and early warning module: Generates a bridge health status report based on the bridge health status prediction results, including the degree of structural damage, bearing capacity attenuation rate, vibration mode change rate, displacement over-limit risk, fatigue damage accumulation, structural stability index, health status score and early warning level; The process of optimizing the monitoring accuracy module and optimizing the hyperparameters of the Qiaoan fusion network includes the following steps: Step B1: define the search space of the hyperparameters of the Bridge-An fusion network and set the k and r double parameters; Step B2: Calculate the divergence gradient of the hyperparameters of the Bridge fusion network; Step B3: Combine the divergence gradient and use the k and r double parameters to update the hyperparameters of the Qiaoan fusion network; Step B4: Set the convergence conditions. If the convergence conditions are met, stop the optimization; otherwise, return to step B2 to continue updating the hyperparameters of the BridgeAn fusion network.

2. A multi-modal bridge vibration monitoring system according to claim 1, characterized in that: The bridge monitoring module generates the bridge health status prediction results, which specifically includes the following steps: Step S1: Establish the topological structure of multimodal bridge data, capture the spatial dependency, temporal dependency and modal dependency of bridge sensors, and construct a sensor interaction graph; Step S2: construct a multi-layer adaptive gated fusion mechanism according to the sensor interaction graph; use the multi-layer adaptive gated fusion mechanism to perform feature fusion and generate spectrum domain fusion features; Step S3: restore missing data of the spectral domain fusion features, and use a multi-head attention mechanism to optimize feature fusion to generate complete bridge data; Step S4: Combine the modal recovery loss function and the cross entropy loss function to construct a bridge health classifier; input the complete bridge data into the bridge health classifier to generate a bridge health status prediction result.

3. A multi-modal bridge vibration monitoring system according to claim 2, characterized in that: Step S2 specifically includes: using the sensor interaction graph to calculate the Laplacian matrix, perform feature decomposition, and obtain low-frequency features and high-frequency features; constructing a multi-layer adaptive gated fusion mechanism through a three-layer gating mechanism, dynamic attention weighting, and gradient constraints; and using a multi-layer adaptive gated fusion mechanism to fuse low-frequency features and high-frequency features to generate spectral domain fusion features.

4. The multi-modal bridge vibration monitoring system according to claim 1, characterized in that: The search space in step B1 includes learning rate, momentum coefficient, regularization parameter and batch size; k, r dual parameters: k controls the smoothness of the change of the hyperparameters of the Bridge Fusion Network, and r controls the sensitivity of the adjustment of the hyperparameters of the Bridge Fusion Network.

5. A multi-modal bridge vibration monitoring system according to claim 4, characterized in that: Step B3 specifically includes the following steps: Step B31: Use k,r-exponential gradient to optimize the learning rate and momentum coefficient; Step B32: Optimizing regularization parameters using k,r-generalized regularization; Step B33: Update the batch size using k,r-additive hyperparameters; Step B34: Use Bregman divergence for collaborative optimization.

Images