Intelligent identification method and device for concealed damage of bridge based on three-source data fusion

Abstract

The application discloses a kind of bridge concealed damage intelligent identification method and device based on three-source data fusion, which comprises acquiring bridge three-source time series data collected by ground-based synthetic aperture radar, satellite navigation system and distributed fiber acoustic sensor;Phase synchronization and data fusion are carried out on the three-source time series data to generate a global fusion constraint field;Based on the global fusion constraint field, the damage time series characteristics of the bridge are extracted;The single-feature damage confidence of each damage time series characteristic is calculated, and the comprehensive damage confidence of the bridge unit is obtained by fusion, and the damaged bridge unit is identified based on the comprehensive damage confidence;Based on the damaged bridge unit, the model is corrected and the stiffness parameter is iterated to update the bridge finite element model, which realizes the accurate positioning and quantitative grading of the concealed micro-damage of the bridge, and accurately corrects the finite element model.

Description

Technical Field

[0001] This application relates to the field of structural health monitoring of highway bridges, specifically to an intelligent identification method and device for hidden bridge damage based on three-source data fusion. Background Technology

[0002] As a crucial infrastructure component of transportation networks, highway bridges' structural safety directly impacts traffic stability and public safety. During long-term service, bridge structures are highly susceptible to various types of damage due to vehicle loads, environmental erosion, and material aging. Hidden areas, such as the interior of box girders, pile foundations, and bearing anchorage zones, are often maintenance blind spots due to the difficulty of detection and complex environments. These hidden micro-damages are characterized by weak initial response signals and insidious evolution processes. Failure to identify and intervene in a timely manner can lead to structural stiffness degradation, crack propagation, and even catastrophic structural defects. Therefore, developing high-precision, intelligent hidden micro-damage identification technology has significant practical engineering value for achieving refined operation and maintenance and safety early warning of bridge structures.

[0003] However, existing technologies for identifying hidden micro-damage in bridges still face numerous technical bottlenecks in practical applications, making it difficult to meet the refined monitoring needs of actual engineering projects. These bottlenecks are manifested in the following aspects: On the one hand, most existing bridge damage identification methods are limited to dual-source fusion data from ground-based interferometric synthetic aperture radar (GB-InSAR) and the BeiDou Navigation Satellite System (BDS), with monitoring dimensions mainly relying on surface deformation and point displacement information. Due to the lack of deep sensing methods for distributed strain and vibration modes, existing technologies cannot effectively obtain stress and strain field changes inside box girders and deep within pile foundations, resulting in significant blind spots in the hidden areas inside the structure.

[0004] Secondly, at the data processing level, existing technologies mostly remain at the basic spatiotemporal registration stage, only achieving physical alignment of data, without truly realizing the deep fusion of multi-dimensional temporal characteristics such as deformation, strain, and vibration. This shallow fusion method is difficult to uncover the coupling effect between different physical quantities, and has limited sensitivity and identification ability for weak and hidden damage.

[0005] Thirdly, at the level of finite element model correction, existing technologies mostly adopt traditional strategies such as global parameter updates or uniform mesh refinement, lacking targeted focus on the damaged area. This correction method not only has high computational redundancy and low efficiency, but also easily introduces parameter perturbations in non-damaged areas, causing the corrected model to deviate from the actual physical state, which seriously restricts the accuracy of locating and quantitatively identifying hidden micro-damages.

[0006] The combined effect of the aforementioned limitations results in a significant deficiency in the sensitivity and reliability of existing methods for detecting subtle and hidden damage, which easily leads to the risk of missed or false diagnoses and hinders the intelligent and refined development of bridge structural health monitoring technology. Summary of the Invention

[0007] This application provides an intelligent identification method and device for hidden bridge damage based on three-source data fusion. It can solve the technical problems in the prior art, such as the single perception dimension of the monitoring means and the insufficient depth of multi-source data fusion, which makes it difficult to accurately identify, locate and classify micro-damage in hidden parts of the bridge. In addition, the correction of the existing finite element model is easily affected by local damage, which leads to the inaccuracy of the model calibration.

[0008] In a first aspect, embodiments of this application provide an intelligent identification method for hidden bridge damage based on three-source data fusion, the intelligent identification method for hidden bridge damage based on three-source data fusion comprising: Acquire three-source time-series data of the bridge collected by ground-based synthetic aperture radar, satellite navigation system and distributed fiber optic acoustic sensors, including bridge deformation time-series data, displacement time-series data, strain time-series data and vibration spectrum data; Phase synchronization and data fusion are performed on the three-source time-series data to generate a global fusion constraint field; Based on the global fusion constraint field, the damage time series characteristics of the bridge are extracted, including the correlation strain time series kurtosis, deformation spatiotemporal joint anomaly index and vibration frequency domain energy skewness index. Calculate the single-feature damage confidence of each damage time-series feature, and fuse them to obtain the comprehensive damage confidence of the bridge unit. Based on the comprehensive damage confidence, identify the damaged bridge unit. The bridge finite element model is updated by modifying the model and iterating the stiffness parameters based on the damaged bridge elements.

[0009] Secondly, embodiments of this application provide a bridge hidden damage device based on three-source data fusion, the bridge hidden damage device based on three-source data fusion includes: The acquisition module is used to acquire three-source time-series data of the bridge collected by ground-based synthetic aperture radar, satellite navigation system and distributed fiber optic acoustic sensor, including bridge deformation time-series data, displacement time-series data, strain time-series data and vibration spectrum data. The fusion module is used to perform phase synchronization and data fusion on the three-source time-series data to generate a global fusion constraint field. The extraction module is used to extract the damage time series features of the bridge based on the global fusion constraint field, including the relevant strain time series kurtosis, deformation spatiotemporal joint anomaly index and vibration frequency domain energy skewness index. The identification module is used to calculate the single-feature damage confidence of each damage time-series feature, and fuse them to obtain the comprehensive damage confidence of the bridge unit, and identify the damaged bridge unit based on the comprehensive damage confidence. The update module is used to perform model correction and stiffness parameter iteration on the bridge finite element model based on the damaged bridge elements, so as to update the bridge finite element model.

[0010] Thirdly, embodiments of this application provide an intelligent identification device for hidden bridge damage based on three-source data fusion. The intelligent identification device for hidden bridge damage based on three-source data fusion includes a processor, a memory, and an intelligent identification program for hidden bridge damage based on three-source data fusion stored in the memory and executable by the processor. When the intelligent identification program for hidden bridge damage based on three-source data fusion is executed by the processor, it implements the steps of the intelligent identification method for hidden bridge damage based on three-source data fusion as described in any of the above claims.

[0011] The beneficial effects of the technical solutions provided in this application include: By acquiring three-source time-series data of bridges collected through ground-based synthetic aperture radar, satellite navigation systems, and distributed fiber optic acoustic sensors, including bridge deformation time-series data, displacement time-series data, strain time-series data, and vibration spectrum data, phase synchronization and data fusion of the three-source time-series data are performed to generate a global fusion constraint field. Based on the global fusion constraint field, the damage time-series features of the bridge are extracted, including the relevant strain time-series kurtosis, the spatiotemporal joint anomaly index of deformation, and the vibration frequency domain energy skewness index. The single-feature damage confidence of each damage time-series feature is calculated and fused to obtain the comprehensive damage confidence of the bridge element, and the damaged bridge element is identified based on the comprehensive damage confidence. Based on the damaged bridge element, the bridge finite element model is corrected and the stiffness parameters are iterated to update the bridge finite element model. This solves the technical problems in related technologies such as the insufficient ability of a single data source to perceive hidden bridge damage, the difficulty in deep fusion of multi-source heterogeneous data, and the inaccuracy of finite element model correction.

[0012] This application constructs a three-source monitoring system by integrating planar deformation acquired by ground-based synthetic aperture radar, absolute displacement acquired by satellite navigation system, and deep strain and vibration spectrum acquired by distributed fiber optic acoustic sensors. This system overcomes the physical limitations of traditional dual-source monitoring, which cannot reach the interior of box girders and the deep layers of pile foundations, and achieves deep stress-strain field perception of the entire structure and hidden areas. By performing phase synchronization and deep fusion of the three-source time-series data, a global time-series deep fusion constraint field is generated. Based on this, multi-dimensional damage time-series characteristics such as strain time-series kurtosis, deformation spatiotemporal joint anomaly index, and vibration frequency domain energy skewness index are extracted, fully exploring the physical coupling effect between deformation, strain, and vibration, greatly enhancing the ability to capture early hidden micro-damage and the anti-interference capability. The application abandons the traditional global uniform correction strategy of finite element model, and only performs high-order element replacement and stiffness parameter iteration on damage-sensitive areas. This variable complexity correction mechanism not only significantly reduces computational redundancy, but also effectively avoids parameter disturbance in non-damaged areas, ensuring the physical authenticity of finite element model calibration and accurate quantification of hidden damage. Attached Figure Description

[0013] Figure 1 This is a flowchart illustrating an embodiment of the intelligent identification method for hidden bridge damage based on three-source data fusion according to this application. Figure 2 This is a schematic diagram of the functional modules of the intelligent identification device for hidden bridge damage based on three-source data fusion in this application; Figure 3 This is a schematic diagram of the hardware structure of the intelligent identification device for hidden bridge damage based on three-source data fusion involved in the embodiments of this application. Detailed Implementation

[0014] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0015] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0016] In one aspect, embodiments of this application provide an intelligent identification method for hidden bridge damage based on three-source data fusion.

[0017] In one embodiment, reference is made to Figure 1 , Figure 1This is a flowchart illustrating the first embodiment of the intelligent identification method for hidden bridge damage based on three-source data fusion according to this application. Figure 1 As shown, the intelligent identification method for hidden bridge damage based on three-source data fusion includes: Step S1: Acquire three-source time-series data of the bridge collected by ground-based synthetic aperture radar, satellite navigation system and distributed fiber optic acoustic sensor, including bridge deformation time-series data, displacement time-series data, strain time-series data and vibration spectrum data.

[0018] Specifically, this embodiment uses ground-based synthetic aperture radar GB-InSAR to acquire global deformation time-series data of the bridge surface (specifically including global time-series deformation curvature and phase characteristic data of the bridge), and uses BeiDou satellite navigation system BDS to collect displacement time-series data of key control points (specifically including three-dimensional displacement time-series and displacement rate characteristic data of key points). Distributed fiber optic acoustic sensors DAS are deployed along the entire length of the main beam, piers and pile foundations to acquire strain time-series data and vibration spectrum data of the entire structure, thereby forming a four-dimensional bridge three-source time-series dataset containing deformation, displacement, strain and vibration.

[0019] Due to environmental interference during data acquisition, the original three-source time-series data often contains anomalous noise and data gaps. To ensure the accuracy of subsequent feature extraction, this embodiment adopts a denoising strategy that combines anomaly removal based on the 3σ criterion with neighborhood mean imputation.

[0020] For the bridge deformation time series data, displacement time series data, strain time series data, and vibration spectrum data, calculate their time series mean μ and standard deviation σ respectively; based on the 3σ criterion, identify and remove data located in the interval [μ]. Outlier data points outside [3σ, μ+3σ] are removed to eliminate the masking effect of sudden strong noise on weak damage signals. For data gaps caused by outlier removal, linear interpolation using the temporal mean method of the preceding and following neighborhoods is used to fill the gaps, ensuring temporal continuity. This denoising strategy smooths random noise while preserving the nonlinear and non-stationary weak response characteristics caused by hidden micro-damage to the greatest extent, making it suitable for the complex service environment of bridge structures.

[0021] Step S2: Perform phase synchronization and data fusion on the three-source time series data to generate a global fusion constraint field.

[0022] In one embodiment, step S2 specifically includes the following steps: Step S201: Transform the coordinates of the three-source time series data into the geodetic coordinate system, map the three-source time series data to the model element mesh of the bridge finite element model through thin plate spline interpolation, and resample to a unified frequency.

[0023] Specifically, this embodiment first unifies the spatial reference and sampling frequency of the three-source time-series data. Using the CGCS2000 National Geodetic Coordinate System as the sole spatial reference, coordinate transformation is performed on the areal deformation data, point displacement data, and linear strain and vibration data to eliminate coordinate system differences between different sensors.

[0024] A unified spatial grid architecture is established based on the unit grid of the bridge finite element model. The BDS monitoring points of the Beidou satellite navigation system are used as high-precision spatial control points. The thin plate spline interpolation TPS algorithm is used to accurately map the planar deformation data and the linear strain and vibration data to the corresponding grid units of the finite element model, thereby realizing the precise matching and one-to-one correspondence of the three sources of data in spatial location.

[0025] After completing the spatial grid mapping, the three-source time series data are further resampled to a preset sampling frequency, such as 1Hz. This completes the preprocessing alignment of the data in both time and space dimensions, laying the data foundation for the subsequent construction of a global time series deep fusion constraint field.

[0026] Step S202: Based on the displacement time series data, rearrange and phase correct the deformation time series data, strain time series data and vibration spectrum data.

[0027] It is worth noting that after completing the spatial dimension mesh mapping, this embodiment further performs time dimension synchronization and alignment on the three-source time series data. Given that the BeiDou Navigation Satellite System (BDS) displacement time series data has the characteristics of stable sampling and clear physical meaning, it is selected as the reference time series, and the deformation time series data collected by the ground-based synthetic aperture radar (GB-InSAR), the strain time series data collected by the distributed fiber acoustic sensor (DAS), and the vibration spectrum data are used as the time series to be synchronized.

[0028] The Dynamic Time Warping (DTW) algorithm is used to calculate the minimum cumulative distance between each timing sequence to be synchronized and the reference timing sequence, and the optimal warping path that minimizes this distance is extracted. Based on the optimal warping path, automatic optimization is performed to establish a one-to-one mapping relationship between different timing data points, thereby effectively eliminating phase misalignment caused by sensor sampling trigger delay, transmission jitter and small deviations in sampling rate.

[0029] Based on the calculated optimal regularization path, the GB-InSAR deformation time series, DAS strain time series and vibration spectrum are rearranged and phase corrected to achieve strict alignment and phase synchronization of the three source signals on the time axis, completely eliminate time offset error, and ensure the instantaneous consistency of physical state during subsequent feature extraction.

[0030] Step S203: Construct a three-dimensional state vector based on the corrected deformation time series data, strain time series data and vibration spectrum data, iteratively update the three-dimensional state vector based on Bayesian adaptive filtering and map it to the bridge global space to generate the global fusion constraint field.

[0031] For example, after completing spatial grid mapping and time phase correction, a three-dimensional state vector X is constructed based on the corrected deformation time series data, strain time series data, and vibration spectrum data. This state vector is defined as X=[d,ε,v]T, where d represents the corrected deformation time series data feature quantity, ε represents the corrected strain time series data feature quantity, and v represents the corrected vibration spectrum data feature quantity. By uniformly encoding the corrected multi-source heterogeneous time series data into a state vector, the purity and physical consistency of the input data are ensured.

[0032] A Bayesian adaptive filtering algorithm is used to iteratively update the three-dimensional state vector, establishing a system model that includes state equations and observation equations, with corrected three-source time-series data as the observation input. Utilizing Bayes' theorem, within each time step, the prior state estimate based on historical data is recursively fused with the current observation value, adaptively correcting the covariance matrix of system noise and observation noise, thereby solving for the optimal posterior state estimate at the current time step. This adaptive mechanism effectively suppresses environmental interference and ensures the stability of the state vector.

[0033] Finally, the filtered and corrected three-dimensional state vector is mapped to the global finite element model element mesh space to generate a global fusion constraint field. This constraint field combines the global deformation evolution law with the deep strain vibration characteristics. The global fusion positioning accuracy is better than 0.1 mm, providing a high-confidence data benchmark for subsequent extraction of high-fidelity damage time series features.

[0034] It is worth noting that the bridge finite element model in this embodiment is based on the bridge design drawings, material physical parameters, and boundary constraints. It is a structural benchmark model established using the general finite element method. The key parts such as the box girder, pile foundation, and anchorage zone are finely meshed. Through modal analysis and static and dynamic calculations, the three-dimensional response benchmark values ​​of the structure under environmental excitation are obtained, providing a benchmark basis for subsequent damage identification and model correction.

[0035] Step S3: Based on the global fusion constraint field, extract the damage time sequence features of the bridge, including the correlation strain time sequence kurtosis, deformation spatiotemporal joint anomaly index and vibration frequency domain energy skewness index.

[0036] In one embodiment, the strain time-series data in the global fusion constraint field is decomposed and transformed to obtain the relevant strain time-series kurtosis:

[0037] in, For the relevant strain time series kurtosis, Let μ be the real-time strain data of the i-th sampling point, and μ be the strain time-series mean. This represents the number of sampling points in the strain time series. The correlation weight is usually determined by the autocorrelation coefficient of the signal, and higher weight is given to periodic recurring shocks.

[0038] Interpretive, Correlated Strain Time-series Kurtosis (CSTK) is used to characterize the nonlinear pulse characteristics of strain time series captured by distributed fiber acoustic sensors (DAS). Under healthy bridge conditions, the stress response of the bridge exhibits an approximately normal distribution, with strain time series data fluctuating smoothly around the mean. The CSTK value remains within the baseline threshold range, demonstrating good steady-state characteristics.

[0039] When hidden damage such as microcracks occurs in the web, bottom plate, or pile foundation of a box girder, the stress concentration effect will cause transient pulse-like anomalies in the local strain field, and the kurtosis will deviate significantly from the reference threshold, which can enable the early identification of weak hidden damage.

[0040] Spatiotemporal correlation analysis was performed on the deformation time-series data in the global fusion constraint field to obtain the joint spatiotemporal anomaly index of deformation:

[0041] Wherein, DSAI is the deformation spatiotemporal joint anomaly index. This represents the temporal data of the planar fused deformation field; x represents the lateral spatial coordinates of the bridge structure, and y represents the longitudinal spatial coordinates of the bridge structure. Let V be the temporal variance of the deformation field.

[0042] The explanatory Deformation Spatio-temporal Anomaly Index (DSAI) is constructed by coupling abrupt changes in the spatial gradient and temporal fluctuations of the deformation field in a planar manner. It is used to quantitatively reflect the degree of distortion in the local deformation of bridge structures. Under healthy structural conditions, the structural stiffness is uniformly distributed, the deformation field is smooth and continuous, and the DSAI maintains a constant low level. When stiffness degradation occurs in key parts such as the web of the box girder, supports, or piers, the redistribution of internal forces in the structure leads to discontinuous abrupt changes (drastic increase in spatial gradient) and non-stationary fluctuations (increased temporal variance) in the local deformation field. This causes the DSAI value to jump significantly and exceed the healthy baseline threshold, thereby enabling precise location of hidden stiffness degradation areas.

[0043] Frequency domain energy statistics are performed on the vibration spectrum data in the global fusion constraint field to obtain the vibration frequency domain energy skewness index:

[0044] Wherein, VSEPI is the vibration frequency domain energy skewness index. Let k be the frequency of the k-th frequency point in the vibration spectrum. The frequency of the vibration spectrum is the weighted average frequency in the frequency domain. Ek represents the percentage of power spectral density at the k-th frequency point, Ehigh represents the total vibration energy of the preset high-frequency band, Etotal represents the total vibration energy of the entire frequency band, and N represents the number of discrete points in the frequency domain.

[0045] Interpretively, the Vibration Spectral Energy Partiality Index (VSEPI) is constructed by coupling the frequency domain distribution skewness characteristics of high-frequency vibration signals from distributed fiber optic acoustic sensors (DAS) with the proportion of high-frequency energy. It is used to sensitively detect structural changes in key force-transmitting connections such as bridge anchorage zones, supports, and expansion joints. Under healthy structural conditions, vibration energy exhibits an approximately symmetrical distribution in the frequency domain, with a low proportion of high-frequency energy, and the VSEPI value remains stable. When prestress relaxation occurs in the anchorage zone, or when supports develop voids or hidden cracks, the contact nonlinearity at the structural connection interface increases, causing frictional slippage and collision effects. This leads to a shift of vibration energy to higher frequencies, resulting in a significant distortion in the frequency domain distribution. Consequently, the VSEPI value jumps sharply and exceeds the healthy baseline threshold, thus providing an early warning of hidden damage to the connection points.

[0046] Step S4: Calculate the single-feature damage confidence of each damage time-series feature, and fuse them to obtain the comprehensive damage confidence of the bridge unit, and identify the damaged bridge unit based on the comprehensive damage confidence.

[0047] It is worth noting that this embodiment further constructs a quantitative mapping mechanism from single-source sensitive features to global damage assessment. The single-feature damage confidence scores of the correlated strain time-series kurtosis (CSTK), the deformation spatiotemporal joint anomaly index (DSAI), and the vibration frequency domain energy skewness index (VSEPI) are calculated separately. Then, considering the differences in the sensitivity of each damage mode to the structural mechanical performance, the three-source confidence scores are adaptively weighted and fused to obtain the comprehensive damage confidence score at the bridge finite element level, effectively avoiding the risk of misjudgment based on a single indicator.

[0048] In one embodiment, based on the relevant strain time-series kurtosis, the deformation spatiotemporal joint anomaly index, and the vibration frequency domain energy skewness index, the corresponding single-feature confidence scores are calculated using a pre-calibrated probability mapping model. (i=CSTK, DSAI, VSEPI) in, For single-feature damage confidence, These are the real-time values ​​corresponding to the temporal characteristics of the damage. As a characteristic health benchmark value; This represents the critical value for characteristic damage. The characteristic health baseline value. The statistical mean or long-term monitoring lower limit of the damage time-series characteristics CSTK, DSAI, and VSEPI corresponding to a bridge structure in a healthy state; characteristic damage threshold. The upper limit value of the feature is determined based on finite element simulation and historical measured values ​​under typical damage conditions. The closer it is to 1, the higher the probability of damage occurring at the corresponding location.

[0049] Based on the sensitivity of each damage time-series feature, the corresponding feature sensitivity coefficient is determined, and the comprehensive damage confidence of the bridge unit is obtained through weighted fusion calculation:

[0050] in, The overall damage confidence level for the bridge unit; , Let be the feature sensitivity coefficient, and satisfy . + + =1. The characteristic sensitivity coefficient is obtained by pre-calibrating the sensitivity of each damage time sequence characteristic to the response of structural stiffness degradation, and is used to quantify the sensitivity of strain, deformation and vibration characteristics under different damage modes.

[0051] Furthermore, the method of identifying damaged bridge units based on comprehensive damage confidence includes: if the comprehensive damage confidence is less than a preset first threshold, the corresponding bridge unit is determined to be intact; if the comprehensive damage confidence is greater than or equal to the first threshold and less than a preset second threshold, the corresponding bridge unit is determined to be slightly damaged; if the comprehensive damage confidence is greater than or equal to the second threshold and less than a preset third threshold, the corresponding bridge unit is determined to be moderately damaged; and if the comprehensive damage confidence is greater than or equal to the third threshold, the corresponding bridge unit is determined to be severely damaged. The damage level determination formula is:

[0052] in, Damage level of bridge units, The first threshold, The second threshold, This is the third threshold.

[0053] This embodiment employs an improved density clustering algorithm for finite element meshes, classifying damage levels based on comprehensive damage confidence. Through spatial neighborhood coupling analysis, this algorithm effectively eliminates isolated noise points and false anomalies, achieving precise localization of hidden micro-damage at the element level. Specifically, for bridge finite element mesh-level identification scenarios, this algorithm constructs a dual-dimensional input of spatial coordinates and comprehensive damage confidence, using confidence to guide clustering and focus on high-suspect areas. It defines neighborhoods based on finite element mesh topology, including only physically connected elements in the same segment and with the same stress attributes, eliminating non-physical connection errors. A comprehensive damage confidence threshold replaces the quantity threshold to determine core points, accurately capturing early, weak, and sparse damage. Spatial coupling verification of CSTK, DSAI, and VSEPI features is added to eliminate interference points with single-feature anomalies, reducing the false alarm rate. Finally, the clustering results are directly mapped to finite element numbers, outputting a structured localization result of bridge element number-location-level, realizing the engineering implementation of damage identification.

[0054] Step S5: Based on the damaged bridge elements, perform model correction and stiffness parameter iteration on the bridge finite element model to update the bridge finite element model.

[0055] In one embodiment, step S5 specifically includes the following steps: Step S501: Construct damage sensitivity factors based on the damage time sequence characteristics of each damaged bridge unit, and determine whether to use the corresponding damaged bridge unit as a damage sensitivity unit based on the damage sensitivity factors.

[0056] Specifically, the damage time-series characteristics corresponding to each damaged bridge unit are numerically normalized. These damaged bridge units include those with minor, moderate, and severe damage levels. Damage sensitivity factors are then constructed based on the normalized damage time-series characteristics.

[0057] in, Let e ​​be the damage sensitivity factor of the damaged bridge element. , , These are the feature weight coefficients, , , This is the normalized value of the damage time sequence feature corresponding to the e-th damaged bridge unit; Set the sensitive area threshold to Damage sensitivity factors of each damaged bridge unit With the sensitive threshold as Perform a comparison; if ≥ , it is determined that the bridge unit is a damage-sensitive unit, which is included in the subsequent finite element model updating scope; otherwise, it is determined as a non-damage-sensitive unit, its unit attributes remain unchanged and it is excluded from the updating calculation. Through this dual-threshold judgment mechanism, the redundant calculation of massive non-damage areas is effectively eliminated, and the interference of parameter disturbance in non-damage areas on the updating results is avoided, ensuring the targeting and efficiency of model updating.

[0058] Step S502: Replace the finite element model unit corresponding to the damage-sensitive unit with a high-order high-precision unit, and keep the non-damage-sensitive unit as a low-order conventional unit unchanged, so as to establish a variable-complexity finite element updating model:

[0059] wherein, is the complexity of the finite element model unit, is the high-order high-precision unit, is the low-order conventional unit, is the damage sensitivity factor of the e-th finite element unit, is the damage sensitivity domain threshold.

[0060] Herein, the high-order high-precision unit generally refers to a second-order or higher-order solid unit, such as a 10-node tetrahedron or 8-node shell unit, which has high resolution for bending deformation and stress gradient. The low-order conventional unit generally refers to a first-order unit, such as a 4-node shell unit or 6-node tetrahedral unit, which has a small amount of calculation but weak ability to fit nonlinear deformation. When SFe≥SFth, it is determined that the unit is located in the potential damage area, and it is replaced from the low-order conventional unit to the high-order high-precision unit to improve the simulation accuracy of the local stress-strain field; when SFe<SFth, the low-order conventional unit is retained to control the overall calculation scale.

[0061] Step S503: Construct the mapping relationship between the stiffness of the damage-sensitive unit and the response of the global fusion constraint field, and with the optimal response matching degree as the goal, iteratively update the stiffness parameters of the finite element model unit corresponding to the damage-sensitive unit to complete the calibration of the bridge finite element model.

[0062] Specifically, the mapping relationship between the stiffness of the damage-sensitive unit and the response of the global fusion constraint field is constructed by the mixed kernel response surface method:

[0063] wherein, is the predicted response of the response surface of the global fusion constraint field, ( ) is the Gaussian radial basis function, , b, and c are the response surface coefficients respectively, is the stiffness sample point, and K is the stiffness parameter vector of the damage-sensitive unit; The iteration is performed with the goal of maximizing the matching coefficient between the predicted response of the global fusion constrained field response surface and the measured response of the global fusion constrained field.

[0064] Where R is the response matching coefficient. For the measured response of the fully integrated constraint field, For covariance, For variance; When the response matching coefficient R is greater than or equal to a preset response matching coefficient threshold, the bridge finite element model calibration is considered complete. In this embodiment, based on engineering experience and numerical experimental verification, the response matching coefficient threshold is set to 0.98; that is, when R ≥ 0.98, it is determined that the predicted response of the finite element model and the measured response of the global time-series deep fusion constraint field achieve a high degree of agreement, iterative convergence is achieved, and model correction is completed. At this time, the stiffness distribution of the finite element model can truly reflect the actual physical state of the bridge structure, effectively eliminating the interference of theoretical model errors on damage identification and positioning accuracy, and ensuring the reliability of the evaluation results.

[0065] In this embodiment, guided by the location results of hidden micro-damage, a three-source fusion damage-sensitive domain is constructed. The traditional global uniform mesh correction strategy is abandoned, and a targeted model update is implemented. Only the damage-sensitive element is replaced by a high-order, high-precision element, while the healthy element remains unchanged as a low-order, conventional element. Thus, while ensuring computational efficiency, a variable-complexity finite element correction model that balances accuracy and efficiency is constructed.

[0066] Building upon this foundation, a hybrid kernel response surface methodology is employed to construct a nonlinear mapping relationship between the stiffness parameters of damage-sensitive elements and the response of the global temporal-series deep fusion constraint field. Using the optimal matching degree between the model-predicted response and the measured response as the objective function, the stiffness parameters of the damaged elements are iteratively updated until the response matching coefficient reaches a preset threshold, thus completing model calibration. This process effectively eliminates the interference of initial theoretical model errors on damage identification and localization, achieving a high degree of isomorphism between the structural physical state and the numerical model, and significantly improving the accuracy and quantification reliability of hidden damage identification at the element level.

[0067] As a preferred implementation, this embodiment also provides visualized result output and decision support functions. After completing global damage identification and model calibration, the system automatically generates a bridge damage location distribution map. This map is displayed based on the overlay of the corrected finite element mesh, intuitively showing the spatial distribution, boundary range, and severity of damaged units in the form of heat maps or contour lines, realizing visualized damage management. Simultaneously, a damage level assessment report is generated, classifying damage into mild, moderate, and severe levels, and providing corresponding safety warning suggestions. Furthermore, a model correction and calibration report is output, recording the results of damage-sensitive unit screening, the iteration history of stiffness parameters, and response matching coefficients, verifying the convergence and accuracy of model correction. Through the above multi-dimensional data results, accurate and reliable decision-making data support is provided to bridge maintenance units, guiding subsequent targeted inspection and reinforcement operations.

[0068] The intelligent identification method for hidden bridge damage based on three-source data fusion provided in this application first breaks through the limitations of traditional single-dimensional monitoring by fusing planar deformation from GB-InSAR, absolute displacement from BDS, and deep strain and vibration from DAS to construct a multi-dimensional sensing system with a unified spatiotemporal reference across the entire domain. Second, it utilizes Bayesian adaptive filtering to generate a deep fusion constraint field across the entire temporal domain and innovatively extracts the relevant strain temporal kurtosis (STK), the deformation spatiotemporal joint anomaly index (DSAI), and the vibration frequency domain energy skewness index (VSEPI), achieving sensitive detection of hidden microcracks, stiffness degradation, and loose connections. Finally, it constructs a variable-complexity finite element correction model based on damage sensitivity factors and uses a hybrid kernel response surface method to target and update the stiffness parameters of damaged elements, eliminating the interference of theoretical model errors on the identification results. This solution solves the technical problems of large blind spots in bridge damage perception, difficulty in identifying weak damage, and high model correction redundancy in existing technologies, achieving millimeter-level precision in element-level positioning and quantitative assessment, providing accurate and reliable full-process technical support for early detection and early warning of hidden bridge damage.

[0069] Secondly, embodiments of this application also provide an intelligent identification device for hidden bridge damage based on three-source data fusion.

[0070] In one embodiment, reference is made to Figure 2 , Figure 2 This is a functional module diagram of an embodiment of the intelligent identification device for hidden bridge damage based on three-source data fusion, as described in this application. Figure 2 As shown, the intelligent identification device for hidden bridge damage based on three-source data fusion includes: The acquisition module is used to acquire three-source time-series data of the bridge collected by ground-based synthetic aperture radar, satellite navigation system and distributed fiber optic acoustic sensor, including bridge deformation time-series data, displacement time-series data, strain time-series data and vibration spectrum data. The fusion module is used to perform phase synchronization and data fusion on the three-source time-series data to generate a global fusion constraint field. The extraction module is used to extract the damage time series features of the bridge based on the global fusion constraint field, including the relevant strain time series kurtosis, deformation spatiotemporal joint anomaly index and vibration frequency domain energy skewness index. The identification module is used to calculate the single-feature damage confidence of each damage time-series feature, and fuse them to obtain the comprehensive damage confidence of the bridge unit, and identify the damaged bridge unit based on the comprehensive damage confidence. The update module is used to perform model correction and stiffness parameter iteration on the bridge finite element model based on the damaged bridge elements, so as to update the bridge finite element model.

[0071] Furthermore, in one embodiment, the fusion module is also used for: The coordinates of the three-source time series data are transformed into the geodetic coordinate system, and the three-source time series data are mapped to the model element mesh of the bridge finite element model through thin plate spline interpolation, and then resampled to a unified frequency. Based on the displacement time series data, the deformation time series data, strain time series data and vibration spectrum data are rearranged and phase corrected; A three-dimensional state vector is constructed based on the corrected deformation time series data, strain time series data and vibration spectrum data. The three-dimensional state vector is iteratively updated based on Bayesian adaptive filtering and mapped to the bridge global space to generate the global fusion constraint field.

[0072] Furthermore, in one embodiment, the extraction module is also used for: The strain time series data in the global fusion constraint field are decomposed and transformed to obtain the relevant strain time series kurtosis:

[0073] in, For the relevant strain time series kurtosis, Let μ be the real-time strain data of the i-th sampling point, and μ be the strain time-series mean. This represents the number of sampling points in the strain time series. Relevance weights; Spatiotemporal correlation analysis was performed on the deformation time-series data in the global fusion constraint field to obtain the joint spatiotemporal anomaly index of deformation:

[0074] Wherein, DSAI is the deformation spatiotemporal joint anomaly index. This represents the temporal data of the planar fused deformation field; x represents the lateral spatial coordinates of the bridge structure, and y represents the longitudinal spatial coordinates of the bridge structure. The time series variance of the deformation field; Frequency domain energy statistics are performed on the vibration spectrum data in the global fusion constraint field to obtain the vibration frequency domain energy skewness index:

[0075] Wherein, VSEPI is the vibration frequency domain energy skewness index. Let k be the frequency of the k-th frequency point in the vibration spectrum. The frequency of the vibration spectrum is the weighted average frequency in the frequency domain. Ek represents the percentage of power spectral density at the k-th frequency point, Ehigh represents the total vibration energy of the preset high-frequency band, Etotal represents the total vibration energy of the entire frequency band, and N represents the number of discrete points in the frequency domain.

[0076] Furthermore, in one embodiment, the identification module is also used for: Based on the relevant strain time-series kurtosis, deformation spatiotemporal joint anomaly index, and vibration frequency domain energy skewness index, the corresponding single-feature confidence scores are calculated using a pre-calibrated probability mapping model. (i=CSTK, DSAI, VSEPI) in, For single-feature damage confidence, These are the real-time values ​​corresponding to the temporal characteristics of the damage. As a characteristic health benchmark value; This is the critical value for characteristic damage; Based on the sensitivity of each damage time-series feature, the corresponding feature sensitivity coefficient is determined, and the comprehensive damage confidence of the bridge unit is obtained through weighted fusion calculation:

[0077] in, The overall damage confidence level for the bridge unit; , Let be the feature sensitivity coefficient, and satisfy . + + =1.

[0078] Furthermore, in one embodiment, the identification module is also used for: If the overall damage confidence level is less than a preset first threshold, the corresponding bridge unit is determined to be of good condition. If the overall damage confidence level is greater than or equal to the first threshold and less than the preset second threshold, then the corresponding bridge unit is determined to be of the minor damage level. If the overall damage confidence level is greater than or equal to the second threshold and less than the preset third threshold, then the corresponding bridge unit is determined to be of moderate damage level. If the overall damage confidence level is greater than or equal to the third threshold, the corresponding bridge unit is determined to be of the severe damage level.

[0079] Furthermore, in one embodiment, the update module is also used to: Damage sensitivity factors are constructed based on the damage time sequence characteristics of each damaged bridge unit, and the corresponding damaged bridge unit is determined as a damage sensitivity unit based on the damage sensitivity factors. The finite element model elements corresponding to the damage-sensitive elements are replaced with high-order, high-precision elements to establish a variable-complexity finite element correction model. A mapping relationship between the stiffness of the damage-sensitive element and the response of the global fusion constraint field is constructed. With the goal of optimal response matching, the stiffness parameters of the finite element model elements corresponding to the damage-sensitive element are iteratively updated to complete the calibration of the bridge finite element model.

[0080] Furthermore, in one embodiment, the update module is also used to: The damage time sequence characteristics corresponding to each damaged bridge unit are numerically normalized. Damage sensitivity factors are constructed based on the normalized damage time sequence characteristics:

[0081] in, Let e ​​be the damage sensitivity factor of the damaged bridge element. , , These are the feature weight coefficients, , , This is the normalized value of the damage time sequence feature corresponding to the e-th damaged bridge unit; If the damage sensitivity factor is greater than or equal to the preset sensitivity threshold, then the corresponding damage bridge unit will be used as the damage sensitivity unit.

[0082] Furthermore, in one embodiment, the update module is also used to: The mapping relationship between the stiffness of the damage-sensitive element and the response of the global fusion constraint field is constructed using the hybrid kernel response surface method:

[0083] in, To predict the response of a globally integrated constrained field response surface, ( ) is a Gaussian radial basis function. b and c are response surface coefficients, respectively. Here, K represents the stiffness sample points, and K is the stiffness parameter vector of the damage-sensitive element. The iteration is performed with the goal of maximizing the matching coefficient between the predicted response of the global fusion constrained field response surface and the measured response of the global fusion constrained field.

[0084] Where R is the response matching coefficient. For the measured response of the fully integrated constraint field, For covariance, For variance; When the response matching coefficient is greater than the preset response matching coefficient threshold, the bridge finite element model calibration is considered complete.

[0085] The functions of each module in the above-mentioned intelligent identification device for hidden bridge damage based on three-source data fusion correspond to the steps in the above-mentioned intelligent identification method for hidden bridge damage based on three-source data fusion. Their functions and implementation processes will not be described in detail here.

[0086] Thirdly, this application provides an intelligent identification device for hidden bridge damage based on three-source data fusion. The intelligent identification device for hidden bridge damage based on three-source data fusion can be a personal computer (PC), laptop computer, server, or other device with data processing capabilities.

[0087] Reference Figure 3 , Figure 3 This is a schematic diagram of the hardware structure of an intelligent bridge hidden damage identification device based on three-source data fusion, as described in an embodiment of this application. In this embodiment, the intelligent bridge hidden damage identification device based on three-source data fusion may include a processor, a memory, a communication interface, and a communication bus.

[0088] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interface.

[0089] The communication interface includes input / output (I / O) interfaces, physical interfaces, and logical interfaces. These interfaces facilitate interconnection of internal components within the intelligent bridge hidden damage identification device based on three-source data fusion, and also enable interconnection between the intelligent bridge hidden damage identification device and other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.

[0090] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0091] The processor can be a general-purpose processor, which can call a smart bridge hidden damage identification program based on three-source data fusion stored in memory and execute the smart bridge hidden damage identification method based on three-source data fusion provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the smart bridge hidden damage identification program based on three-source data fusion is called can refer to the various embodiments of the smart bridge hidden damage identification method based on three-source data fusion in this application, and will not be repeated here.

[0092] Those skilled in the art will understand that Figure 3 The hardware structure shown does not constitute a limitation of this application and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0093] It should be noted that the sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0094] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0095] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0096] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0097] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0098] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0099] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for intelligent identification of hidden bridge damage based on three-source data fusion, characterized in that, The intelligent identification method for hidden bridge damage based on three-source data fusion includes: Acquire three-source time-series data of the bridge collected by ground-based synthetic aperture radar, satellite navigation system and distributed fiber optic acoustic sensors, including bridge deformation time-series data, displacement time-series data, strain time-series data and vibration spectrum data; Phase synchronization and data fusion are performed on the three-source time-series data to generate a global fusion constraint field; Based on the global fusion constraint field, the damage time series characteristics of the bridge are extracted, including the correlation strain time series kurtosis, deformation spatiotemporal joint anomaly index and vibration frequency domain energy skewness index. Calculate the single-feature damage confidence of each damage time-series feature, and fuse them to obtain the comprehensive damage confidence of the bridge unit. Based on the comprehensive damage confidence, identify the damaged bridge unit. Based on the damaged bridge elements, the bridge finite element model is modified and the stiffness parameters are iterated to update the bridge finite element model. Specifically, based on the global fusion constraint field, the damage temporal features of the bridge are extracted, including: The strain time series data in the global fusion constraint field are decomposed and transformed to obtain the relevant strain time series kurtosis: in, For the relevant strain time series kurtosis, Let μ be the real-time strain data of the i-th sampling point, and μ be the strain time-series mean. This represents the number of sampling points in the strain time series. Relevance weights; Spatiotemporal correlation analysis was performed on the deformation time-series data in the global fusion constraint field to obtain the joint spatiotemporal anomaly index of deformation: Wherein, DSAI is the deformation spatiotemporal joint anomaly index. This represents the temporal data of the planar fused deformation field; x represents the lateral spatial coordinates of the bridge structure, and y represents the longitudinal spatial coordinates of the bridge structure. The time series variance of the deformation field; Frequency domain energy statistics are performed on the vibration spectrum data in the global fusion constraint field to obtain the vibration frequency domain energy skewness index: Wherein, VSEPI is the vibration frequency domain energy skewness index. Let k be the frequency of the k-th frequency point in the vibration spectrum. The frequency of the vibration spectrum is the weighted average frequency in the frequency domain. The proportion of the power spectral density at the k-th frequency point. The total vibrational energy in the preset high-frequency band, Let N be the total vibration energy across the entire frequency band, and N be the number of discrete points in the frequency domain.

2. The intelligent identification method for hidden bridge damage based on three-source data fusion as described in claim 1, characterized in that, Phase synchronization and data fusion are performed on the three-source time series data to generate a global fusion constraint field, including: The coordinates of the three-source time series data are transformed into the geodetic coordinate system, and the three-source time series data are mapped to the model element mesh of the bridge finite element model through thin plate spline interpolation, and then resampled to a unified frequency. Based on the displacement time series data, the deformation time series data, strain time series data and vibration spectrum data are rearranged and phase corrected; A three-dimensional state vector is constructed based on the corrected deformation time series data, strain time series data and vibration spectrum data. The three-dimensional state vector is iteratively updated based on Bayesian adaptive filtering and mapped to the bridge global space to generate the global fusion constraint field.

3. The intelligent identification method for hidden bridge damage based on three-source data fusion as described in claim 1, characterized in that, The calculation of the single-feature damage confidence of each damage time-series feature and the fusion of these features to obtain the comprehensive damage confidence of the bridge unit includes: Based on the relevant strain time-series kurtosis, deformation spatiotemporal joint anomaly index, and vibration frequency domain energy skewness index, the corresponding single-feature damage confidence is calculated using a pre-calibrated probability mapping model. ( =CSTK、DSAI、VSEPI) in, For single-feature damage confidence, These are the real-time values ​​corresponding to the temporal characteristics of the damage. As a characteristic health benchmark value; This is the critical value for characteristic damage; Based on the sensitivity of each damage time-series feature, the corresponding feature sensitivity coefficient is determined, and the comprehensive damage confidence of the bridge unit is obtained through weighted fusion calculation: in, The overall damage confidence level for the bridge unit; , Let be the feature sensitivity coefficient, and satisfy . + + =1.

4. The intelligent identification method for hidden bridge damage based on three-source data fusion as described in claim 1, characterized in that, Damaged bridge elements are identified based on comprehensive damage confidence, including: If the overall damage confidence level is less than a preset first threshold, the corresponding bridge unit is determined to be of good condition. If the overall damage confidence level is greater than or equal to the first threshold and less than the preset second threshold, then the corresponding bridge unit is determined to be of the minor damage level. If the overall damage confidence level is greater than or equal to the second threshold and less than the preset third threshold, then the corresponding bridge unit is determined to be of moderate damage level. If the overall damage confidence level is greater than or equal to the third threshold, the corresponding bridge unit is determined to be of the severe damage level.

5. The intelligent identification method for hidden bridge damage based on three-source data fusion as described in claim 1, characterized in that, Based on the damaged bridge elements, the bridge finite element model is modified and its stiffness parameters are iterated to update the bridge finite element model, including: Damage sensitivity factors are constructed based on the damage time sequence characteristics of each damaged bridge unit, and the corresponding damaged bridge unit is determined as a damage sensitivity unit based on the damage sensitivity factors. The finite element model elements corresponding to the damage-sensitive elements are replaced with high-order, high-precision elements to establish a variable-complexity finite element correction model. A mapping relationship between the stiffness of the damage-sensitive element and the response of the global fusion constraint field is constructed. With the goal of optimal response matching, the stiffness parameters of the finite element model elements corresponding to the damage-sensitive element are iteratively updated to complete the calibration of the bridge finite element model.

6. The intelligent identification method for hidden bridge damage based on three-source data fusion as described in claim 5, characterized in that, Damage sensitivity factors are constructed based on the damage time sequence characteristics of each damaged bridge unit. The determination of whether to designate a corresponding damaged bridge unit as a damage-sensitive unit is based on these damage sensitivity factors, including: The damage time sequence characteristics corresponding to each damaged bridge unit are numerically normalized. Damage sensitivity factors are constructed based on the normalized damage time sequence characteristics: in, Let e ​​be the damage sensitivity factor of the damaged bridge element. , , These are the feature weight coefficients, , , This is the normalized value of the damage time sequence feature corresponding to the e-th damaged bridge unit; If the damage sensitivity factor is greater than or equal to the preset sensitivity threshold, then the corresponding damage bridge unit will be used as the damage sensitivity unit.

7. The intelligent identification method for hidden bridge damage based on three-source data fusion as described in claim 5, characterized in that, A mapping relationship is constructed between the stiffness of the damage-sensitive element and the response of the global fusion constraint field. With the goal of optimal response matching, the stiffness parameters of the finite element model elements corresponding to the damage-sensitive element are iteratively updated to complete the bridge finite element model calibration, including: The mapping relationship between the stiffness of the damage-sensitive element and the response of the global fusion constraint field is constructed using the hybrid kernel response surface method: in, To predict the response of a globally integrated constrained field response surface, ( ) represents the Gaussian radial basis functions. , c and c are response surface coefficients, respectively. Here, K represents the stiffness sample points, and K is the stiffness parameter vector of the damage-sensitive element. The iteration is performed with the goal of maximizing the matching coefficient between the predicted response of the global fusion constraint field response surface and the measured response of the global fusion constraint field. Where R is the response matching coefficient. For the measured response of the fully integrated constraint field, For covariance, For variance; When the response matching coefficient is greater than the preset response matching coefficient threshold, the bridge finite element model calibration is considered complete.

8. A bridge hidden damage device based on three-source data fusion, characterized in that, The bridge hidden damage device based on three-source data fusion includes: The acquisition module is used to acquire three-source time-series data of the bridge collected by ground-based synthetic aperture radar, satellite navigation system and distributed fiber optic acoustic sensor, including bridge deformation time-series data, displacement time-series data, strain time-series data and vibration spectrum data. The fusion module is used to perform phase synchronization and data fusion on the three-source time-series data to generate a global fusion constraint field. The extraction module is used to extract the damage time series characteristics of the bridge based on the global fusion constraint field, including the correlation strain time series kurtosis, deformation spatiotemporal joint anomaly index and vibration frequency domain energy skewness index. The identification module is used to calculate the single-feature damage confidence of each damage time-series feature, and fuse them to obtain the comprehensive damage confidence of the bridge unit, and identify the damaged bridge unit based on the comprehensive damage confidence. The update module is used to perform model correction and stiffness parameter iteration on the bridge finite element model based on the damaged bridge elements, so as to update the bridge finite element model. The extraction module is further used for: The strain time series data in the global fusion constraint field are decomposed and transformed to obtain the relevant strain time series kurtosis: in, For the relevant strain time series kurtosis, Let μ be the real-time strain data of the i-th sampling point, and μ be the strain time-series mean. The number of sampling points in the strain time series. Relevance weights; Spatiotemporal correlation analysis was performed on the deformation time-series data in the global fusion constraint field to obtain the joint spatiotemporal anomaly index of deformation: Wherein, DSAI is the deformation spatiotemporal joint anomaly index. This represents the temporal data of the planar fused deformation field; x represents the lateral spatial coordinates of the bridge structure, and y represents the longitudinal spatial coordinates of the bridge structure. The time series variance of the deformation field; Frequency domain energy statistics are performed on the vibration spectrum data in the global fusion constraint field to obtain the vibration frequency domain energy skewness index: Wherein, VSEPI is the vibration frequency domain energy skewness index. Let k be the frequency of the k-th frequency point in the vibration spectrum. The frequency of the vibration spectrum is the weighted average frequency in the frequency domain. E represents the proportion of the power spectral density at the k-th frequency point. high E is the total vibrational energy in the preset high-frequency band. total Let N be the total vibration energy across the entire frequency band, and N be the number of discrete points in the frequency domain.

9. An intelligent identification device for hidden bridge damage based on three-source data fusion, characterized in that, The intelligent identification device for hidden bridge damage based on three-source data fusion includes a processor, a memory, and an intelligent identification program for hidden bridge damage based on three-source data fusion stored in the memory and executable by the processor. When the intelligent identification program for hidden bridge damage based on three-source data fusion is executed by the processor, it implements the steps of the intelligent identification method for hidden bridge damage based on three-source data fusion as described in any one of claims 1 to 7.

Images