A bridge main structure health monitoring and abnormality identification method

Abstract

The present invention belongs to the field of bridge main structure monitoring, and specifically discloses a bridge main structure health monitoring and anomaly identification method. Vibration data is collected by arranging a vibration sensor array in the load distribution area of the bridge span structure, and the wind-induced component correction of the vibration signal is performed in combination with real-time wind speed data, thereby achieving accurate collection of vibration data required for bridge structure health monitoring. At the same time, when identifying bridge structure anomalies based on the vibration signal after wind-induced vibration correction, a vibration energy propagation path map is constructed by jointly analyzing the phase and amplitude of the vibration wave data. On this basis, anomalies are identified in combination with the frequency concentration and amplitude mutation of the node. On the one hand, non-structural vibration interference is effectively eliminated through propagation path modeling; on the other hand, the energy state change of the node in the propagation path is utilized to achieve accurate identification of the energy anomaly propagation position, thereby improving the accuracy of damage location.

Description

Technical Field

[0001] The present invention belongs to the field of bridge main structure monitoring, and specifically discloses a bridge main structure health monitoring and abnormality identification method. Background Art

[0002] As key nodes in transportation networks, bridges play a vital role in ensuring road capacity and regional connectivity. However, over the course of long-term service, their structures are susceptible to performance degradation due to multiple factors, such as vehicle dynamic loads and material aging, impacting their safety, durability, and operational reliability. Therefore, it is essential to conduct structural health monitoring of bridges during their service.

[0003] During the service life of bridges, vehicle loads, as the primary external excitation source, induce periodic or transient vibration responses in the span structure. This response not only reflects the dynamic characteristics of the structure but can also manifest as shifts in energy propagation paths, frequency dispersion, and abnormal amplitude fluctuations due to localized stiffness variations, loose connections, or altered mass distribution. Therefore, monitoring methods based on vibration signals have been widely adopted in bridge structural health monitoring systems due to their high sensitivity and adaptability.

[0004] Currently, bridge structural health monitoring technologies based on vibration signals are widely used in engineering practice. For example, Chinese Invention Patent Publication No. CN116842348A proposes an artificial intelligence-based bridge health monitoring system. This system periodically collects bridge vibration data caused by vehicle traffic, identifies peak points and performs amplitude fitting on the vibration signals, and extracts an initial abnormality index to assess abnormal trends in the structural operating state. Furthermore, this method decomposes the vibration data and combines it with the predicted vibration responses of each vehicle at different time points to dynamically correct the initial abnormality level, obtaining a more representative target abnormality level. This allows for the effective screening of abnormal data and identification of structural conditions, improving the reliability and accuracy of the monitoring results.

[0005] While the aforementioned methods are somewhat effective in identifying vibration anomalies under vehicle load excitation, they fail to fully consider the impact of wind loads on bridge vibration responses during modeling and data analysis. For long-span bridges or those located in open areas, wind loads are a significant external excitation source that cannot be ignored. They can cause significant aerodynamic vibrations, which in turn interfere with vibration monitoring data. Failure to identify and separate these wind-induced vibration components can lead to misinterpretations of vibration anomaly characteristics, compromising the accuracy of structural damage identification. Summary of the Invention

[0006] To this end, one purpose of an embodiment of the present application is to provide a method for monitoring the health and identifying abnormalities of a bridge main structure by placing sensors on the bridge span structure to collect vibration signals, and correcting the wind-induced vibration components of the vibration signals in combination with real-time wind speed data, thereby accurately identifying abnormal structural areas based on the corrected vibration data, effectively solving the problems existing in the prior art.

[0007] The purpose of the present invention can be achieved by the following technical solutions: A method for monitoring the health of a bridge main structure and identifying abnormalities, comprising the following steps: (1) dividing the load distribution area on the bridge span structure to perform wind field simulation, analyzing the wind pressure sensitivity based on the analysis, and setting the vibration sensor layout spacing based on the wind pressure sensitivity to form a vibration sensor array.

[0008] (2) Vibration wave data is collected through a vibration sensor array, and wind speed is collected using an anemometer installed on the bridge deck. A unified timestamp is added to the vibration wave data and wind speed data to generate a synchronized monitoring data set.

[0009] (3) Compare the wind speed in the synchronous monitoring data set with the safe wind speed. When the wind speed is less than the safe value, retain the original vibration wave data. When the wind speed reaches the safe value, perform wind-induced vibration correction on the vibration wave data.

[0010] (4) Analyze the phase and amplitude characteristics of the corrected vibration wave data, analyze the propagation path of the vibration energy in the bridge structure, and generate a vibration energy propagation path map.

[0011] (5) Calculate the frequency concentration and amplitude mutation of each node in the vibration energy propagation path map, and compare them with the historical health status database to identify structural abnormal areas.

[0012] Combining all the above technical solutions, the positive effects of the present invention are as follows: 1. The present invention collects vibration data by deploying a vibration sensor array in the load distribution area of the bridge span structure, and corrects the wind-induced component of the vibration signal in combination with real-time wind speed data, thereby achieving accurate collection of vibration data required for bridge structure health monitoring, ensuring data reliability, significantly reducing the misjudgment rate of abnormal features, and improving damage identification accuracy and monitoring system stability.

[0013] The present invention analyzes the wind pressure sensitivity by conducting wind field simulation on the load distribution area of the bridge span structure, and accordingly densely deploys sensors in wind pressure-sensitive areas, while appropriately reducing the deployment density in areas with less wind pressure impact, thereby achieving spatially differentiated and optimized configuration of the sensor network, avoiding sensor redundancy or insufficient coverage caused by traditional uniform deployment, improving system deployment efficiency, and providing a data foundation with reasonable spatial distribution and strong time synchronization for subsequent vibration signal correction based on wind speed data, thereby improving the accuracy and stability of the structural health monitoring system.

[0014] When identifying bridge structure anomalies based on vibration signals after wind-induced vibration correction, the present invention jointly analyzes the phase and amplitude of vibration wave data, extracts the timing correlation and energy transfer relationship between sensor nodes, and constructs a vibration energy propagation path map with physical significance. On this basis, the frequency concentration and amplitude mutation of each node are combined to identify anomalies. On the one hand, non-structural vibration interference is effectively eliminated through propagation path modeling, focusing on the real structural response area and reducing the misjudgment rate; on the other hand, the energy state changes of the nodes in the propagation path are utilized to realize the accurate identification of the energy abnormality propagation position, thereby improving the accuracy of damage location. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention is further described with reference to the accompanying drawings. However, the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative effort.

[0016] Figure 1 This is a diagram of the steps for implementing the method of the present invention.

[0017] Figure 2 It is a schematic diagram of the directional energy propagation and outgoing energy propagation of the corresponding nodes in the vibration energy propagation path diagram in the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] See also Figure 1 As shown, the present invention proposes a method for health monitoring and abnormality identification of a bridge main structure, comprising the following steps: (1) dividing the load distribution area on the bridge span structure to perform wind field simulation, analyzing the wind pressure sensitivity based on the load distribution area, and setting the vibration sensor layout spacing based on the wind pressure sensitivity to form a vibration sensor array.

[0020] It should be noted that the main bridge structure referred to in this invention specifically refers to the span structure, i.e., the primary load-bearing portion of a bridge that spans over obstacles. It directly bears the loads of vehicles, pedestrians, and the environment, and effectively transmits these loads to the piers and foundations. A typical span structure includes key components such as the main beam, deck, diaphragm, and supports.

[0021] By deploying sensors on the structure to collect vibration signals, the dynamic response characteristics of the bridge under actual loads can be truly reflected. The acquired vibration data is highly representative and highly correlated with the structural status, providing a high-quality data foundation for subsequent anomaly identification and health assessment.

[0022] As a preferred implementation of the above steps, the load distribution area is divided on the bridge span structure for wind field simulation, and the wind pressure sensitivity is analyzed based on this, including the establishment of a three-dimensional bridge model and wind field simulation, wherein the establishment of the three-dimensional bridge structure model is implemented as follows: the load distribution area is divided on the bridge span structure according to the design load distribution diagram of the bridge.

[0023] It's important to understand that while the span structure is the primary component of a bridge bearing external loads, the actual load bearing process is uneven across its various regions, and not all parts directly participate in the primary load transmission. Therefore, based on the bridge's design load distribution diagram, this paper identifies and delineates load distribution areas within the span structure—critical areas within the structure where load-bearing effects are significant and stress and strain are concentrated.

[0024] Deploying sensors within this area and collecting vibration signals more effectively captures the bridge's key dynamic response characteristics under stimuli such as vehicle traffic. The resulting vibration data is more structured and representative, improving the accuracy of anomaly identification and monitoring efficiency. This approach avoids ineffective deployment of monitoring points in non-critical areas, optimizes the allocation of monitoring resources, and enhances the system's sensitivity to structural state changes and the reliability of its judgments.

[0025] A high-precision three-dimensional bridge model is established based on the bridge structure design drawings and geometric parameters, and meshing is completed within the load distribution area.

[0026] Applied to the above scheme, the bridge geometric parameters include but are not limited to overall dimensions: total length of the bridge, main span length, bridge deck width, cross-sectional shape and dimensions: main beam cross-sectional form, cross-sectional height, thickness of top plate, bottom plate, and web, material properties: elastic modulus, Poisson's ratio, density,

[0027] Detailed structure: diaphragm arrangement, support position and type, stiffening rib setting, connection details: welded or bolted connection nodes, joint treatment method.

[0028] The above-mentioned establishment of a high-precision three-dimensional bridge model based on the bridge structure design drawings and geometric parameters is intended to ensure that the model shape is consistent with the actual bridge structure.

[0029] Furthermore, completing meshing within the load distribution area can achieve local mesh refinement in that area, thereby improving the spatial resolution of the numerical model. This helps to more accurately simulate the dynamic response behavior of the bridge structure under external excitation and improve the ability to capture the vibration signal propagation path and local vibration characteristics.

[0030] As a further preferred implementation of the above steps, the wind field simulation is implemented as follows: long-term historical meteorological data, including wind speed, wind direction, and seasonal variation characteristics, are obtained from the meteorological station in the area where the bridge is located, and the inlet wind speed boundary conditions and outlet pressure boundary conditions are set in combination with the local topographic information, in accordance with the basic principles of fluid mechanics, to ensure that the simulation environment reflects the actual working conditions as realistically as possible.

[0031] Arranging the historical wind speed data in ascending order covers the entire process from conventional wind loads to extreme wind loads, which helps to identify the differences in structural responses under different wind speeds. Based on this order, the wind loads under different wind speeds are simulated in sequence on the established three-dimensional bridge model and its set boundary conditions to obtain the wind pressure distribution diagram of the load distribution area under different wind speed conditions.

[0032] It should be added that wind pressure refers to the pressure generated on the surface of the structure due to the action of wind. Commonly used wind field simulation methods include CFD simulation, which can describe in detail how wind interacts with the bridge structure and produces complex wind pressure distribution. By solving the Navier-Stokes equations or other turbulence models, the wind speed distribution at each point on the bridge surface can be obtained, and then the corresponding wind pressure distribution can be calculated according to the Bernoulli equation.

[0033] The above wind pressure distribution diagram connects the wind pressure values of adjacent measuring points into a continuous regional distribution through the spatial interpolation method, which vividly expresses the spatial heterogeneity, gradient change and local extreme value distribution of wind pressure in the load distribution area.

[0034] According to the wind pressure distribution diagram of the load distribution area under multi-wind speed simulation, the areas with average wind pressure lower than the preset threshold are screened out and defined as wind pressure non-sensitive areas, and the remaining areas are classified as wind pressure sensitive areas.

[0035] It should be noted that the above-mentioned threshold reflects the critical sensitivity level of the bridge structure to wind loads under multi-speed simulation conditions. Specifically, this threshold is used to differentiate the strength of the force response of different areas of the bridge span structure surface under wind loads and serves as a criterion for determining whether a particular area is experiencing significant wind-induced excitation.

[0036] The setting of preset thresholds can be based on the structural characteristics of the bridge and engineering experience. Specifically, long-term meteorological data can be used to establish typical wind environment conditions, and combined with CFD simulation to obtain the average wind pressure values of each area under various wind speeds. The wind pressure distribution of all grid points is statistically analyzed, and its mean or percentile (such as the 75th percentile) is selected as the threshold reference value.

[0037] In the above, areas where the average wind pressure is lower than the preset threshold are defined as wind pressure-insensitive areas. The basis for this is that wind pressure is the main external excitation source that causes wind-induced vibration. The greater the wind pressure, the stronger the wind-induced excitation of the structure, and the more significant the corresponding vibration response. When the average wind pressure in a certain area is lower than the set threshold, it indicates that the wind load in this area under typical wind conditions is relatively weak, the probability of wind-induced vibration is low, and its vibration signal is mainly affected by vehicle traffic load and has low wind correlation. Therefore, such areas can be reasonably classified as wind pressure-insensitive areas, so that they can be treated differently in subsequent vibration data analysis and sensor deployment, thereby improving the pertinence and effectiveness of the monitoring system.

[0038] For the wind pressure sensitive areas, the corresponding wind pressure values under different wind speed simulations are extracted, and the wind pressure response curve is drawn with wind speed as the horizontal axis and wind pressure value as the vertical axis.

[0039] The overall change rate is extracted from the wind pressure response curve as the wind pressure growth amplitude, and the initial wind pressure value at the lowest wind speed is extracted as the basic wind pressure.

[0040] It should be noted that the overall rate of change extracted from the wind pressure response curve is used to characterize the overall growth trend of wind pressure with increasing wind speed, reflecting the dynamic evolution characteristics of wind pressure borne by the bridge structure under different wind speed excitations.

[0041] Since the wind pressure response curve usually exhibits nonlinear characteristics, its changing trend is not a straight line with a constant slope. For example, in order to accurately obtain the change rate, the wind pressure response curve can be fitted to construct a functional relationship between wind pressure and wind speed. The average change rate is calculated based on the fitting function, that is, the ratio of the wind pressure change to the wind speed change is obtained within the wind speed range as a quantitative indicator of the overall growth trend.

[0042] The basic wind pressure and wind pressure growth amplitude are normalized respectively, and then the product of the normalized basic wind pressure and wind pressure growth amplitude in the wind pressure sensitive area is taken as the wind pressure sensitivity.

[0043] In this paper, base wind pressure reflects the initial stress level of the bridge structure under the lowest wind speed conditions, while the wind pressure growth amplitude characterizes the overall change trend of wind pressure as wind speed increases. To eliminate differences in dimensions and numerical ranges between the two parameters and improve comparability across different regions, these two parameters are normalized so that their values are uniformly mapped to the [0, 1] interval.

[0044] On this basis, the normalized base wind pressure is multiplied by the wind pressure growth rate to construct a comprehensive wind pressure sensitivity index. This index comprehensively considers the initial stress state of the structure at low wind speeds and the development trend of wind pressure response at high wind speeds, and can scientifically and reasonably quantify the sensitivity of different areas to wind loads.

[0045] Its mathematical properties indicate that wind pressure sensitivity increases significantly only when both the base wind pressure and the wind pressure growth rate are high. If either parameter is low, the combined index value is also suppressed, thus demonstrating a synergistic constraint effect on the wind pressure response intensity. This construction mechanism conforms to the principle of structural response superposition and the characteristics of nonlinear excitation response, and can more realistically reflect the stress behavior of bridge structures in complex wind environments.

[0046] Furthermore, forming a vibration sensor array by differentially setting the spacing of vibration sensors based on wind pressure sensitivity includes the following: arranging vibration sensors at set spacings in wind pressure-insensitive areas to meet the needs of basic vibration status monitoring in the area.

[0047] For wind pressure sensitive areas, wind pressure sensitivity is introduced as a layout density adjustment factor, and then the layout spacing of the corresponding area is obtained by multiplying the set spacing and the layout density adjustment factor.

[0048] The density adjustment factor is defined as 1 minus the wind pressure sensitivity value.

[0049] It needs to be explained that vibration sensors are deployed at set intervals in wind pressure non-sensitive areas, and wind pressure sensitivity is introduced in wind pressure sensitive areas for denser deployment. This is because the vibration energy caused by wind in wind pressure sensitive areas is stronger and the frequency is higher, so higher spatial resolution is required to capture vibration signals. Under the premise of ensuring overall monitoring efficiency, the monitoring resolution of areas with strong wind-induced vibration response can be improved, resource allocation can be optimized, and system costs can be reduced.

[0050] The present invention analyzes the wind pressure sensitivity by conducting wind field simulation on the load distribution area of the bridge span structure, and accordingly densely deploys sensors in wind pressure-sensitive areas, while appropriately reducing the deployment density in areas with less wind pressure impact, thereby achieving spatially differentiated and optimized configuration of the sensor network, avoiding sensor redundancy or insufficient coverage caused by traditional uniform deployment, improving system deployment efficiency, and providing a data foundation with reasonable spatial distribution and strong time synchronization for subsequent vibration signal correction based on wind speed data, thereby improving the accuracy and stability of the structural health monitoring system.

[0051] (2) Vibration wave data is collected through a vibration sensor array, and wind speed is collected using an anemometer installed on the bridge deck. A unified timestamp is added to the vibration wave data and wind speed data to generate a synchronized monitoring data set.

[0052] Applied to the above steps, the synchronous monitoring data set is specifically formed as follows: ultrasonic anemometers are installed symmetrically on both sides of the bridge deck to measure the wind speed in real time.

[0053] The measured wind speed and vibration sensor data are time-synchronized through a unified clock source to generate a synchronized monitoring dataset with a timestamp.

[0054] The wind speed in the synchronous monitoring data set is compared with the safe wind speed. When the wind speed is lower than the safe value, the original vibration wave data is retained. When the wind speed reaches the safe value, the vibration wave data is corrected for wind-induced vibration.

[0055] It should be understood that when the wind speed is lower than the safe value, it indicates that the current wind load has little excitation on the bridge structure, and the wind-induced vibration response is negligible or secondary. At this time, the vibration of the bridge is mainly caused by structural excitations such as vehicle loads and environmental perturbations, which has good physical representativeness and stability. There is no need to perform wind-induced vibration correction, and the original vibration wave data can be directly used to meet the signal authenticity requirements of structural health monitoring.

[0056] When the wind speed reaches a safe value, the wind load becomes an important external excitation source that cannot be ignored, which may cause significant wind-induced vibration. Such vibration components mixed in the original vibration signal may mask the true structural response characteristics, leading to misjudgment of abnormality identification or distortion of modal parameter estimation. Therefore, it is necessary to separate and correct the wind-induced vibration components of the original vibration signal to extract effective vibration information reflecting the state of the structure itself.

[0057] Specifically, the safe wind speed can be set based on the critical wind speed parameter determined during the structural design and wind resistance performance assessment of the bridge, which is usually derived from the wind resistance design standard of the bridge.

[0058] In the optional implementation of the above steps, wind-induced vibration correction refers to the following process: frequency domain decomposition is performed on the vibration wave data collected by each vibration sensor in the load distribution area to separate the low-frequency structural vibration component and the high-frequency wind-induced vibration component.

[0059] It should be pointed out that the decomposition of vibration wave data into low-frequency structural vibration components and high-frequency wind-induced vibration components is based on the differential characteristics of the natural frequency of the bridge structure and the wind-induced vibration excitation frequency band. The low-frequency structural vibration component mainly reflects the intrinsic response of the structure caused by traffic loads, foundation excitation, etc., while the high-frequency wind-induced vibration component mainly comes from aerodynamic excitation under the action of wind loads, and manifests as wide-band fluctuations.

[0060] The amplitude envelope of the high-frequency wind-induced vibration component is extracted, where the amplitude envelope reflects the temporal variation of the wind-induced vibration intensity. The real-time wind speed data recorded by the anemometer is then time-aligned with the high-frequency wind-induced vibration amplitude envelope of the corresponding period.

[0061] A quantitative mapping relationship between wind speed and wind-induced vibration amplitude is established based on historical data.

[0062] In the above implementation example, the quantitative mapping of wind speed and wind-induced vibration amplitude can construct a comparison data set of wind speed and wind-induced vibration amplitude based on historical monitoring data, and use regression analysis method to fit the functional relationship between the two, so as to obtain a prediction model that can be used for real-time wind-induced vibration correction.

[0063] The above mapping relationship is used to predict the wind-induced vibration amplitude under the current wind speed conditions, and it is deducted from the original vibration signal as a background interference component.

[0064] In the exemplary application of the above operation, the specific implementation of subtracting from the original vibration signal is: subtracting the estimated wind-induced vibration component from the original signal in the time domain.

[0065] The wind-induced vibration correction of the vibration wave signal collected by the vibration sensor is aimed at eliminating the interference caused by wind load on the vibration monitoring data of the bridge structure and improving the recognition accuracy of the structure's own response component in the vibration signal.

[0066] (4) Analyze the phase and amplitude characteristics of the corrected vibration wave data, analyze the propagation path of the vibration energy in the bridge structure, and generate a vibration energy propagation path map.

[0067] The specific implementation process of the above steps is as follows: (41) According to the actual layout position of each vibration sensor, define the set of sensors directly adjacent to each sensor node to form the adjacent area of the node.

[0068] (42) For each sensor node, the phase difference and amplitude difference of the vibration wave signals after wind-induced vibration correction are extracted from the sensor node and the sensors in the adjacent area at the same time stamp are calculated. If the following conditions are met between a sensor node and a sensor in the adjacent area:

[0069] a) The phase difference is less than the preset phase threshold, indicating that there is an obvious vibration sequence between the two

[0070] b) The amplitude difference is within the set attenuation range, indicating that energy is transferred from one end to the other rather than randomly fluctuating.

[0071] It is determined that there is a vibration energy propagation relationship between the two, and the physical position connection line between the sensor node and its adjacent sensor is marked as a vibration energy propagation segment.

[0072] It should be explained that the propagation of vibration waves in bridge structures follows the classical wave equation. Vibration energy generally propagates along the direction where the structural stiffness is maximum and the mass distribution is most continuous. By constructing adjacent areas for local vibration propagation analysis, compared with directly identifying the global propagation path within the load distribution area, which is computationally intensive and susceptible to noise interference, defining adjacent areas and using local comparison methods can effectively narrow the search space, improve recognition efficiency and robustness, and also be more in line with the local propagation characteristics of actual vibration waves.

[0073] In addition, the vibration energy propagation path analysis is performed in the adjacent area based on the phase difference and amplitude difference of the vibration waves collected by two adjacent vibration sensors. The reason is that the phase difference reflects the time delay characteristics of the vibration energy propagating from one point to another, which can be used to determine the propagation direction of the vibration energy; the amplitude difference reflects the energy attenuation caused by material damping, interface reflection, local defects, etc. during the propagation of the vibration wave in the structure, which is used to verify whether the vibration energy is actually transmitted. The combination of the two can significantly improve the reliability of propagation path identification. In this process, by setting the phase threshold and amplitude difference attenuation range, false propagation paths can be effectively filtered out to avoid misjudgment. Specifically, the phase threshold and amplitude difference attenuation range can be used to simulate the vibration propagation process under typical excitation using a finite element model, and the theoretical phase difference and amplitude difference between different sensors can be extracted as a reference.

[0074] Starting from the new node identified on the vibration energy propagation segment, the analysis is repeated with its adjacent area as the object of analysis (42), and the propagation path is gradually expanded outward until the entire sensor network is covered or the set maximum tracking depth is reached or no new effective propagation segment is found.

[0075] All identified propagation segments are integrated into a complete vibration energy propagation path map, and the propagation direction is marked in the propagation path map.

[0076] The above-mentioned propagation segments based on the identification are continuously extended outward to form a complete propagation path, which reflects the spatial continuity of the structural response.

[0077] (5) Calculate the frequency concentration and amplitude mutation of each node in the vibration energy propagation path map, and compare them with the historical health status database to identify structural abnormal areas.

[0078] In the optional implementation of the above steps, the frequency concentration and amplitude mutation of each node in the vibration energy propagation path map are calculated as follows: the vibration energy propagation path map is marked with nodes, and the vibration wave data at each node is transformed in the frequency domain to convert the time domain signal into a frequency domain representation to obtain the spectrum amplitude distribution.

[0079] Based on the obtained spectrum amplitude distribution, the main frequency components with significant energy contribution are identified, and the ratio of the sum of the energy corresponding to all main frequency points to the total energy of the entire spectrum is calculated as the frequency concentration of the node.

[0080] In the example operation of the above implementation, the main frequency components can be extracted from the spectrum amplitude distribution in the following manner: finding the frequency corresponding to the peak in the spectrum or screening out significant frequency components according to a preset frequency range, such as the natural frequency interval of the bridge structure.

[0081] It should be noted that frequency concentration refers to the energy proportion of the main frequency components in the vibration signal collected at a specific node. A high frequency concentration means that the vibration response at that node is more stable and controlled by the dominant mode of the structure, while a low frequency concentration indicates a more dispersed vibration spectrum. In a healthy state, the main vibration frequencies of a bridge structure are generally stable and concentrated in a few natural frequencies, resulting in a high frequency concentration. When the structure experiences local damage or stiffness changes, new frequency components may be introduced or the energy distribution of the original frequency components may be changed, resulting in a decrease in frequency concentration.

[0082] The vibration wave data at each node is analyzed in time series to extract its instantaneous amplitude sequence, and then the amplitude variation between adjacent time points is calculated to form an amplitude difference sequence.

[0083] On the basis of obtaining the amplitude difference sequence, the average absolute difference value is further calculated as the amplitude mutation degree of the node.

[0084] It should be pointed out that amplitude mutation is used to measure the severity of the change in the amplitude of the vibration signal over time. It reflects whether there is a sudden energy release or absorption phenomenon in the node area. Under normal circumstances, the vibration amplitude change of the bridge structure should be relatively stable, and the amplitude mutation is low. If problems such as crack expansion and support failure occur in a certain area, it may cause significant fluctuations in the local vibration amplitude, which is manifested as an increase in amplitude mutation.

[0085] By combining the above two indicators, the condition of the bridge structure can be evaluated more comprehensively and potential problem areas can be identified in a timely manner.

[0086] In a further optional implementation of the above steps, comparing it with the historical health status database to identify the structural abnormality area includes the following process: Figure 2 As shown in the figure, the node weight factor of each node in the vibration energy propagation path map is constructed based on the number of energy propagation segments pointing to the node and the number of propagation segments starting from the node, combined with normalization processing and weighted fusion strategy.

[0087] It should be noted that in the above Figure 2 The black dots in the represent nodes.

[0088] Specifically, the number of energy propagation segments pointing to a node reflects the receiving capacity of the node in the structural dynamic response, and the number of propagation segments starting from the node reflects the driving effect of the node on the surrounding area.

[0089] In order to eliminate the influence of graph scale and node distribution density when constructing node weight factors using the two quantities, the above two parameters are divided by the corresponding maximum number of incoming propagation segments and the maximum number of outgoing propagation segments in the entire graph to achieve a unified numerical range. Further, an adjustable parameter is introduced to linearly weight the normalized number of incoming and outgoing propagation segments to obtain the node weight factor.

[0090] In an example expression, , where Indicates the node number, , 、 Respectively indicate the direction The number of energy propagation segments of the node, from The number of propagation segments that the node originates from, 、 Respectively represent the maximum number of incoming propagation segments and the maximum number of outgoing propagation segments in the entire graph, Indicates an adjustable parameter, the value is The adjustable parameter can be flexibly set based on actual monitoring objectives and analysis needs to adjust the relative influence of the number of inbound and outbound propagation segments in the node weight factor. When focusing on energy convergence, a higher weight can be given to the inbound propagation segment, i.e., a larger adjustable parameter value can be selected. Conversely, if the focus is on energy dissipation, the adjustable parameter value can be lowered to emphasize the influence of the outbound propagation segment.

[0091] The node weight factor reflects the importance of the node as an energy convergence point. The higher the value, the more critical the node is in the vibration energy propagation path.

[0092] The frequency concentration and amplitude mutation of each node in the current monitoring state are compared with their corresponding benchmark frequency concentration and benchmark amplitude mutation in the historical health status database to obtain the frequency deviation and amplitude deviation.

[0093] The above-mentioned historical health status database stores the baseline frequency concentration and baseline amplitude mutation of each node of the bridge when it is in a known health state. The specific construction method is: after the bridge is completed and initially debugged, comprehensive vibration testing and environmental monitoring are carried out. The vibration signals of each sensor node under various typical working conditions are collected, and the corresponding frequency concentration and amplitude mutation are calculated. These are stored in the database as initial baseline data. The bridge is regularly inspected at preset time intervals, and the latest normal status data is added to the database to ensure that the baseline remains up to date over time.

[0094] Specifically, the frequency deviation can be obtained by taking the difference between the node's reference frequency concentration and the frequency concentration and dividing it by the reference frequency concentration, which reflects the degree of deviation of the node's frequency concentration from the reference;

[0095] The amplitude deviation can be obtained by dividing the difference between the node's amplitude mutation and the benchmark amplitude mutation by the benchmark amplitude mutation, reflecting the degree of deviation of the node's amplitude mutation from the benchmark.

[0096] A preset global warning deviation threshold is introduced, and it is dynamically adjusted according to the weight factor of each node to obtain the warning deviation threshold of each node.

[0097] As a specific implementation example, the calculation expression of the warning deviation threshold of each node is: , where Represents the global warning deviation threshold, Indicates the The node's warning deviation threshold.

[0098] It should be understood that the global warning deviation threshold introduced above refers to the upper limit of the unified benchmark deviation used to determine whether there is abnormal behavior in the structural node.

[0099] An exemplary acquisition method is to set a reasonable deviation tolerance value by referring to relevant industry standards for bridge structure health monitoring.

[0100] From the above expression, we can see that when the weight factor of the node is larger, The smaller it is, the lower the node's warning deviation threshold is relative to the global warning deviation threshold, which means that for nodes with higher importance in the vibration energy propagation path, the upper limit of the allowed deviation is smaller and the abnormality identification standard is more stringent.

[0101] For each node, the calculated frequency deviation and amplitude deviation are compared with the corresponding warning deviation threshold. If any indicator exceeds the corresponding threshold, it is determined that there is a structural abnormality in the area corresponding to the node, and the abnormal area is marked.

[0102] This node anomaly identification mechanism fully considers the topological roles of different sensor nodes in the vibration energy propagation path. By introducing a node weight factor to dynamically adjust the warning deviation threshold, the anomaly judgment criteria not only rely on the degree of deviation of local vibration characteristics but also incorporate the influence weight of the node in the entire structural response network. This method not only improves the accuracy of anomaly identification, but also enhances the physical consistency and engineering interpretability of the judgment results, helping to more effectively identify critical anomaly areas with structural impact.

[0103] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0104] Those skilled in the art will appreciate that the algorithmic steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0105] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0106] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A bridge main structure health monitoring and abnormality identification method, characterized in that: The following steps are involved: (1) The load distribution area on the bridge span structure is divided into wind field simulations, and the wind pressure sensitivity is analyzed based on the wind pressure sensitivity. The vibration sensor layout spacing is set based on the wind pressure sensitivity to form a vibration sensor array; (2) Vibration wave data is collected through a vibration sensor array, and wind speed is collected using an anemometer installed on the bridge deck. A unified time stamp is added to the vibration wave data and wind speed data to generate a synchronized monitoring data set; (3) Compare the wind speed in the synchronous monitoring data set with the safe wind speed. When the wind speed is less than the safe value, retain the original vibration wave data. When the wind speed reaches the safe value, perform wind-induced vibration correction on the vibration wave data. The wind-induced vibration correction process is as follows: Perform frequency domain decomposition on the vibration wave data collected by each vibration sensor in the load distribution area to separate the low-frequency structural vibration component and the high-frequency wind-induced vibration component; Extract the amplitude envelope of the high-frequency wind-induced vibration component and perform time series alignment with the high-frequency wind-induced vibration amplitude envelope of the corresponding period based on the real-time wind speed data recorded by the anemometer; Establish a quantitative mapping relationship between wind speed and wind-induced vibration amplitude based on historical data; The above mapping relationship is used to predict the wind-induced vibration amplitude under the current wind speed conditions, and it is deducted from the original vibration signal as a background interference component; (4) Analyze the phase and amplitude characteristics of the corrected vibration wave data, analyze the propagation path of the vibration energy in the bridge structure, and generate a vibration energy propagation path map; (5) Calculate the frequency concentration and amplitude mutation of each node in the vibration energy propagation path map, and compare them with the historical health status database to identify structural abnormal areas.

2. The bridge main structure health monitoring and abnormality identification method according to claim 1, characterized in that: The load distribution area is divided on the bridge span structure to perform wind field simulation, and the wind pressure sensitivity analysis is based on this, which includes the establishment of a three-dimensional bridge model and wind field simulation. The three-dimensional bridge structure model is established as follows: Divide the load distribution area on the bridge span structure according to the design load distribution diagram of the bridge; A high-precision three-dimensional bridge model is established based on the bridge structure design drawings and geometric parameters, and meshing is completed within the load distribution area.

3. The bridge main structure health monitoring and abnormality identification method according to claim 2, characterized in that: The wind field simulation is implemented as follows: Obtain long-term historical meteorological data from meteorological stations in the bridge area, including wind speed, wind direction, and seasonal variation characteristics. Combined with local topographic information, the inlet wind speed boundary conditions and outlet pressure boundary conditions are set. The historical wind speed data are arranged in ascending order, and based on this order, wind loads under different wind speeds are simulated in sequence on the established three-dimensional bridge model and its set boundary conditions to obtain wind pressure distribution diagrams of the load distribution area under different wind speed conditions; Based on the wind pressure distribution diagram of each load distribution area under multi-wind speed simulation, the areas with average wind pressure lower than the preset threshold are screened and defined as wind pressure non-sensitive areas, and the remaining areas are classified as wind pressure sensitive areas; For wind pressure sensitive areas, the corresponding wind pressure values under different wind speed simulations are extracted, and a wind pressure response curve is drawn with wind speed as the horizontal axis and wind pressure value as the vertical axis; The overall change rate is extracted from the wind pressure response curve as the wind pressure growth amplitude, and the initial wind pressure value at the lowest wind speed is extracted as the basic wind pressure; The basic wind pressure and wind pressure growth amplitude are normalized respectively, and then the product of the normalized basic wind pressure and wind pressure growth amplitude in the wind pressure sensitive area is taken as the wind pressure sensitivity.

4. The bridge main structure health monitoring and abnormality identification method according to claim 3 is characterized by: The step of setting the vibration sensor arrangement spacing based on the wind pressure sensitivity to form a vibration sensor array includes the following: Vibration sensors are placed at set intervals in areas not sensitive to wind pressure; For wind pressure sensitive areas, wind pressure sensitivity is introduced as a layout density adjustment factor, and then the layout spacing of the corresponding area is obtained by multiplying the set spacing and the layout density adjustment factor; The density adjustment factor is defined as 1 minus the wind pressure sensitivity value.

5. The bridge main structure health monitoring and abnormality identification method according to claim 1, characterized in that: The synchronous monitoring data set is specifically formed as follows: Ultrasonic anemometers are installed symmetrically on both sides of the bridge deck to measure wind speed in real time; The measured wind speed and vibration sensor data are time-synchronized through a unified clock source to generate a synchronized monitoring dataset with a timestamp.

6. The bridge main structure health monitoring and abnormality identification method according to claim 1, characterized in that: The specific implementation process of step (4) is as follows: (41) According to the actual layout position of each vibration sensor, define the set of sensors directly adjacent to each sensor node to form the adjacent area of the node; (42) For each sensor node, the phase difference and amplitude difference of the vibration wave signals after wind-induced vibration correction are extracted from the sensor node and the sensors in the adjacent area at the same time stamp are calculated. If the following conditions are met between a sensor node and a sensor in the adjacent area: a) The phase difference is less than the preset phase threshold; b) The amplitude difference is within the set attenuation range; It is determined that there is a vibration energy propagation relationship between the two, and the physical position connection line between the sensor node and its adjacent sensor is marked as a vibration energy propagation segment; Starting from the new node identified on the vibration energy propagation segment, the analysis is repeated with its adjacent area as the object (42), and the propagation path is gradually expanded outward until the entire sensor network is covered or the set maximum tracking depth is reached or no new valid propagation segment is found; All identified propagation segments are integrated into a complete vibration energy propagation path map, and the propagation direction is marked in the propagation path map.

7. The bridge main structure health monitoring and abnormality identification method according to claim 1, characterized in that: The specific implementation of calculating the frequency concentration of each node in the vibration energy propagation path map is as follows: The vibration energy propagation path map is marked with nodes, and the vibration wave data at each node is transformed into a frequency domain to convert the time domain signal into a frequency domain representation to obtain the spectrum amplitude distribution; Based on the obtained spectrum amplitude distribution, the main frequency components with significant energy contribution are identified, and the ratio of the sum of the energy corresponding to all main frequency points to the total energy of the entire spectrum is calculated as the frequency concentration of the node.

8. The bridge main structure health monitoring and abnormality identification method according to claim 1, characterized in that: The amplitude mutation degree is calculated as follows: Perform time series analysis on the vibration wave data at each node to extract its instantaneous amplitude sequence, and then calculate the amplitude variation between adjacent time points to form an amplitude difference sequence; On the basis of obtaining the amplitude difference sequence, the average absolute difference value is further calculated as the amplitude mutation degree of the node.

9. The bridge main structure health monitoring and abnormality identification method according to claim 1, characterized in that: Comparing it with the historical health status database and then identifying the structural abnormality area includes the following process: The node weight factor of each node in the vibration energy propagation path map is constructed based on the number of energy propagation segments pointing to the node and the number of propagation segments starting from the node, combining normalization processing and weighted fusion strategy; The frequency concentration and amplitude mutation of each node in the current monitoring state are compared with the corresponding benchmark frequency concentration and benchmark amplitude mutation in the historical health status database to calculate the frequency deviation and amplitude deviation; A preset global warning deviation threshold is introduced and dynamically adjusted according to the weight factor of each node to obtain the warning deviation threshold of each node; For each node, the calculated frequency deviation and amplitude deviation are compared with the corresponding warning deviation threshold. If any indicator exceeds the corresponding threshold, it is determined that there is a structural abnormality in the area corresponding to the node, and the abnormal area is marked.

Images